sign in sign up
<<
Home , Alvarezsauridae, Avisauridae, Baptornis, Cathayornis, Enaliornis, Eoenantiornis

and the Origin of : A Review and Two New Analyses

OM66_FM.indd 1 3/31/09 4:56:43 PM Ornithological Monographs

Editor: John Faaborg

224 Tucker Hall Division of Biological Sciences University of Missouri Columbia, Missouri 65211

Managing Editor: Mark C. Penrose Copy Editor: Richard D. Earles

Authors of this issue: Frances C. James and John A. Pourtless IV

Translation of the abstract by Lisbeth O. Swain

The Ornithological Monographs series, published by the American Ornithologists’ Union, has been established for major papers and presentations too long for inclusion in the Union’s journal, The Auk.

Copying and permissions notice: Authorization to copy article content beyond fair use (as specified in Sections 107 and 108 of the U.S. Copyright Law) for internal or personal use, or the internal or personal use of specific clients, is granted by The Regents of the University of California on behalf of the American Ornithologists’ Union for libraries and other users, provided that they are registered with and pay the specified fee through the Copyright Clearance Center (CCC), www.copyright.com. To reach the CCC’s Customer Service Department, phone 978-750-8400 or write to info@copyright. com. For permission to distribute electronically, republish, resell, or repurpose material, and to purchase article offprints, use the CCC’s Rightslink service, available on Caliber at http://caliber. ucpress.net. Submit all other permissions and licensing inquiries through University of California Press’s Rights and Permissions website, www.ucpressjournals.com/reprintInfo.asp, or via e-mail: [email protected].

Back issues of Ornithological Monographs from earlier than 2007 are available from Buteo Books at http://www.buteobooks.com. Issues from 2008 forward are available from University of California Press at http://www.ucpressjournals.com. For complete abstracting and indexing coverage for Ornithological Monographs, please visit http://www.ucpressjournals.com.

Library of Congress Control Number 2009923609 ISBN: 978-0-943610-85-6 Issued 30 April 2009 Ornithological Monographs, No. 66 viii + 78 pp.

Cover: If core maniraptoran theropod (, , and Oviratorosauria) were actually flightless and flying birds that were more derived towards modern birds than , then the hypothesis that birds are maniraptoran theropod dinosaurs would lose most of its current support, and the would have to be evaluated in the light of at least four other hypotheses (see Fig. 3 on page 6).

Printed by Allen Press on Forest Stewardship Council–certified paper.

© 2009 American Ornithologists’ Union. All rights reserved.

OM66_FM.indd 2 3/31/09 4:56:50 PM Cladistics and : A Review and Two New Analyses

by

Frances C. James1 and John A. Pourtless IV2

1Department of Biological Science, Florida State University, Tallahassee, Florida 32306, USA; and 21327 High Road #K-1, Tallahassee, Florida 32304, USA

ORNITHOLOGICAL MONOGRAPHS NO. 66 PUBLISHED BY THE AMERICAN ORNITHOLOGISTS’ UNION WASHINGTON, D.C. 2009

OM66_FM.indd 3 3/31/09 4:56:51 PM Table of Contents

List of tables and figures...... vi ABSTRACT...... 1 INTRODUCTION...... 2 Definition of Aves...... 4 THE BMT HYPOTHESIS AND ALTERNATIVES ...... 4 The bmt Hypothesis...... 5 The Neoflightless-theropod Hypothesis...... 7 The Early- Hypothesis...... 8 The Crocodylomorph Hypothesis ...... 8 The Hypothesis that Birds Evolved Twice...... 11 CLADISTICS...... 11 METHODS ...... 11 Construction of Matrices ...... 11 Methods of Analysis of the Matrices ...... 14 RESULTS...... 16 Reanalysis of the cnm Matrix ...... 16 Analysis of the New Matrix ...... 18 Alternative Analysis of the New Matrix ...... 23 DISCUSSION ...... 25 Has the bmt Hypothesis Been Tested?...... 25 Unjustifiable assumptions of incorporated into data matrices ...... 25 Inadequate sampling...... 25 Insufficiently rigorous application of cladistic methods...... 26 Insufficient tests of primary homology statements ...... 26 Lack of statistical evaluation...... 27 Verificationist arguments...... 27 Introduction of ad-hoc auxiliary hypotheses...... 27 Evaluation of Alternative Hypotheses...... 28 Are predictions of the BMT hypothesis supported by our results? ...... 28 Are some maniraptorans actually birds more derived than Archaeopteryx? ...... 30 Weakened support for the BMT hypothesis and the neoflightless-theropod hypothesis...... 32 Are alternatives to the BMT hypothesis compatible with the evidence? ...... 35 Conclusions ...... 38 ACKNOWLEDGMENTS...... 39 LITERATURE CITED...... 39 APPENDIX 1: SPECIMENS EXAMINED ...... 52 APPENDIX 2: CHARACTER LIST FOR 242 CHARACTERS, NOTES, AND REFERENCES...... 52 Notes...... 61 APPENDIX 3: ANALYSIS OF CHARACTERS IN THE MANUS, CARPUS, AND TARSUS THAT WERE EXCLUDED FROM THE PRIMARY ANALYSIS OF OUR NEW MATRIX ...... 63 Manus...... 63 Carpus...... 64

v

OM66_FM.indd 5 3/31/09 4:56:51 PM Disappearance of the avian ulnare and development of the pisiform ...... 65 Homology of the proximal carpals in ...... 65 Distribution of proximal carpals in nonmaniraptoran theropods ...... 65 Homology of the radiale within Aves and between birds and theropods ...... 66 Homology of the semilunate carpal within Aves ...... 66 Homology of the semilunate carpal within the and between theropods and birds ...... 67 Identification of distal carpals without knowledge of digital identity ...... 68 Tarsus: Ascending Process of the Astragalus...... 68 APPENDIX 4: DATA MATRIX ...... 69 APPENDIX 5: NOTES ACCOMPANYING MATRIX ON SCORING DECISIONS FOR CERTAIN TAXA ...... 76

LIST OF TABLES

1. Taxa included in our analysis and the principal references used for coding their characters ...... 9 2. Statistics from reanalysis of the 46-taxon × 208-character matrix of Clark et al. (2002) . . 19 3. Statistics from analysis of our 79-taxon × 221-character matrix with lengths of most-parsimonious trees (MPTs) when certain theropod taxa were constrained to be within Aves as defined here ...... 23 4. Statistics from analysis of our 79 taxon × 221 character matrix with lengths of most-parsimonious trees (MPTs) when taxa were constrained to match hypotheses for the origin of birds (BMT = birds are maniraptoran theropod dinosaurs) ...... 23 5. Five predictions of the BMT (birds are maniraptoran theropod dinosaurs) hypothesis . 29 6. Some of the major potential synapomorphies with birds that support the three major alternative hypotheses for the origin of birds ...... 33

LIST OF FIGURES

1. Alternative relationships of birds and selected maniraptorans ...... 4 2. Generally accepted phylogeny of the Archosauria including the topology of the BMT hypothesis ...... 5 3. Five major hypotheses for the origin of birds ...... 6 4. Embryonic bud at (A) 5.5 days of development and (B) 7.5 days of development ...... 13 5. Carpal in the chicken wing bud at (A) 9.5, (B) 11.5, and (C) 12.5 days of development ...... 14 6. (A) Distal and proximal tarsals of , (B) a composite of the London and Berlin specimens of Archaeopteryx, and (C) the Thermopolis specimen of Archaeopteryx . . . . . 15 7. A 50% majority-rule summary of 500 bootstrap replicates of most-parsimonious trees calculated from the matrix reported by Clark et al. (2002) ...... 17 8. As in Figure 7, but pruned to 24 taxa that represent all the major ingroup of Clark et al. (2002) ...... 18 9. A 50% majority-rule summary of 500 bootstrap replicates of most-parsimonious trees calculated from a matrix of 79 taxa including theropods, birds, nontheropod dinosaurs, and other , pruned to 20 taxa ...... 20 10. As in Figure 9, but with also pruned away ...... 20

vi

OM66_FM.indd 6 3/31/09 4:56:51 PM 11. As in Figure 9, pruned to 35 taxa from 79, representing all major proposed sister groups of birds ...... 21 12. As in Figure 11, but with the Maniraptora pruned away ...... 22 13. As in Figure 12, but with both maniraptorans and Longisquama pruned away . . . . . 22 14. A 50% majority-rule summary of 500 bootstrap replicates of most-parsimonious trees calculated from an alternative analysis to the analyses reported in Figures 9–13, using 242 instead of 221 characters, and pruned as in Figure 11 ...... 24 15. of Longisquama insignis as preserved on the counterslab of the specimen (left) and a reconstruction and interpretation of the cranial osteology (right) ...... 37

vii

OM66_FM.indd 7 3/31/09 4:56:51 PM Ornithological Monographs Volume (2009), No. 66, 1–78 © The American Ornithologists’ Union, 2009. Printed in USA.

Cladistics and the Origin of Birds: A Review and Two New Analyses Frances C. James1,3 and John A. Pourtless IV2 1Department of Biological Science, Florida State University, Tallahassee, Florida 32306, USA; and 21327 High Road, no. K-1, Tallahassee, Florida 32304, USA

Abstract.—The hypothesis that birds are maniraptoran theropod dinosaurs (the “BMT hypo­ thesis”) has become widely accepted by both scientists and the general public. Criticism has usu­ ally been dismissed, often with the comment that no more parsimonious alternative has been presented with cladistic methodology. Rather than taking that position, we ask here whether the hypothesis is as overwhelmingly supported as some claim. We reanalyzed a standard matrix of 46 taxa and 208 characters from a recent paper by Clark, Norell, and Makovicky, and we found statistical support for the clades and and for a of birds and maniraptorans. Note, however, that because the matrix contains only birds and theropods, it assumes that the origin of birds lies within the Theropoda. In addition to this problem, Clark et al.’s (2002) matrix contains problematic assumptions of homology, especially in the palate, basipterygoid, manus, carpus, and tarsus. In an attempt to avoid these two major problems and to evaluate the BMT hypothesis and four alternative hypotheses in a comparative phylogenetic framework, we followed the recommendations of Jenner, Kearney, and Rieppel by constructing and analyzing a larger but more conservative matrix. Our matrix includes taxa from throughout the Archosauria. When the ambiguous characters were excluded, parsimony analyses with boot­ strapping and successive pruning retrieved a weak clade of birds and core maniraptorans (ovirap­ torosaurs, troodontids, and dromaeosaurs) that also contained the early archosaur Longisquama and was not unambiguously associated with other theropods. When the ambiguous characters were included but coded as unknown where appropriate, the results were virtually identical. Kishino-Hasegawa tests revealed no statistical difference between the hypothesis that birds were a clade nested within the Maniraptora and the hypothesis that core clades of Maniraptora were actually flying and flightless radiations within the clade bracketed by Archaeopteryx and mod­ ern birds (Aves). Additional statistical tests showed that both the “early-archosaur” and “cro­ codylomorph” hypotheses are at least as well supported as the BMT hypothesis. These results show that Theropoda as presently constituted may not be monophyletic and that the verification­ ist approach of the BMT literature may be producing misleading studies on the origin of birds. Further research should focus on whether some maniraptorans belong within Aves, and whether Aves belongs within Theropoda or is more closely related to another archosaurian taxon. At pres­ ent, uncertainties about the hypothesis that birds are maniraptoran theropods are not receiving enough attention. Received 28 July 2008, accepted 25 January 2009.

Resumen.—La hipótesis de que las aves son dinosaurios maniraptores terópodos (la teoría de BMT) ha sido extensamente aceptada por científicos y el público en general. Criticas en contra de esta hipótesis han sido usualmente rechazadas basandose en el hecho de que una alternativa mas parsimoniosa no ha sido presentada usando metodologia cladística. Nosotros cuestionamos si la hipótesis esta realmente tan bien respaldada como ha sido indicado previamente. Se reanalizó una matriz estándar de 46 taxones y 208 caracteres basados en el artículo publicado por James Clark, Mark Norrell y Peter Makovicky, encontrando resultados estadísticos positivos que respaldan los clados de Coelurosauria y Maniraptoriformes y para los clados de aves y Maniraptora. Sin embargo, es importante notar que ya que la matriz contiene solamente aves y terópodos, esta

3E-mail: [email protected]

Ornithological Monographs, Number 66, pages 1–78. ISBN: 978-0-943610-80-1. © 2009 by The American Ornithologists’ Union. All rights reserved. Please direct all requests for permission to photocopy or reproduce article content through the University of California Press’s Rights and Permissions website, http://www.ucpressjournals.com/reprintInfo.asp. DOI: 10.1525/om.2009.66.1.1. 1

OM66_Text.indd 1 4/3/09 10:20:39 AM 2 ORNITHOLOGICAL MONOGRAPHS NO. 66

asume que el origen de las aves se sitúa dentro del grupo de los terópodos. Adicionalmente a este problema, las matrices de Clark, Norrell y Makovick también contienen problemáticas suposi­ ciones de homologia especialmente en el paladar, basipterygoid, manus, carpus y tarsos. Haciendo un intento en evitar estos dos grandes problemas y de evaluar la teoría de BMT y cuatro hipótesis alternativas en un marco filogenético comparativo, nosotros utilizamos las recomendaciones de Jenner, Kearney, y Rieppel construyendo y analizando una matriz más grande y conservadora. Nuestra muestra incluye diferentes taxones representantes de los arcosaurios. Cuando lo carac­ teres ambiguos fueron excluidos, el análisis de parsimonia con bootstrapping y sucesivas podas produjeron un débil clado de aves y del núcleo de los maniraptores (oviraptóridos troodóntidos dromaeosáuridos) que también contiene al arcosaurios Longisquama y no fue asociado sin ambig­ üedad con otros terópodos. Cuando los caracteres ambiguos fueron incluidos pero codificados como desconocidos en casos apropiados, los resultados fueron virtualmente idénticos. Los resul­ tados de los pruebas Kishino-Hasegawa revelan que no hay una diferencia estadística entre la hipótesis que sustenta que las aves eran un clado perteneciente a los dinosaurious Maniraptores, y la hipótesis que propone que los clados de los dinosaurios Maniraptores eran radiaciones de voladores y no voladores dentro del clado bordeado por Archaeopteryx y aves modernas. Otros análisis estadísticos indican que las hipótesis del “ancestro-arcosaurio” y Crocodylomorpha son tan bien apoyadas como la hipótesis de BMT. Estos resultados demuestran que los terópodos como son constituidos hoy en día pueden no ser monofiléticos, y por lo tanto el enfoque verifica­ dor de la literatura de la BMT puede que produzca resultados erróneos sobre el origen de las aves. Mas estudios deberían enfocarse en descifrar su algunos dinosaurios maniraptores pertenecen a las aves, y si las aves pertenecen a la familia de los terópodos o si están relacionadas mas cercana­ mente a otro taxón de arcosaurios. Hasta el momento, no se le ha prestado mucha atención a las incertidumbres que genera la hipótesis que indica que las aves son maniraptores terópodos.

Introduction argued that Archaeopteryx was primarily and incapable of powered and that the long The current consensus hypothesis about the of dromaeosaurs like were origin of birds, often called “the theropod hy­ “preadapted” for flight. The BMT hypothesis was pothesis,” states that the sister clade of Aves is more explicitly formulated on the basis of cladistic among the five groups of maniraptoran thero­ analyses by Padian (1982), Gauthier and Padian pod dinosaurs (, Troodontidae, (1985), and Gauthier (1986). Gauthier (1986) con­ Dromaeosauridae, , and Ther­ cluded that his analysis of the large group izinosauroidea). The Troodontidae and the Dro­ supported a sister-group relationship maeosauridae, which together constitute the between birds and either the or Deinonychosauria (Colbert and Russell 1969, Gau­ the Dromaeosauridae alone. thier 1986, Norell and Makovicky 2004), are gener­ Since the publication of Gauthier’s influential ally agreed to be the sister clade of Aves. The clade 1986 paper, many more data have become avail­ uniting Deinonychosauria and Aves is sometimes able. New birds are known from labeled (e.g., Turner et al. 2007). Aves is, , , Spain, , therefore, deeply nested within Maniraptora and and especially China (Padian 2004, Zhou 2004, Theropoda. We will call this hypothesis the “BMT and references therein). Remarkable -like hypothesis” (from “birds are maniraptoran thero­ maniraptorans have been found in China, and pod dinosaurs”). By implication, birds are avian others are known from Mongolia and elsewhere maniraptorans, and the Oviraptorosauria, Troo­ (Weishampel et al. 2004, and references therein). dontidae, Dromaeosauridae, Alvarezsauridae, and Structures identified as were reported in Therizinosauroidea are nonavian maniraptorans. the compsognathid in 1996 (for a The BMT hypothesis was derived from of the detailed description, see Currie and Chen 2001), work of (1973, 1975, 1976a, b). Ostrom and similar structures have since been reported in described extensive osteological similarities be­ other theropod taxa (see Weishampel et al. 2004, tween the of the dromaeosaurid Deinon- and references therein). We now know that vari­ ychus and the earliest known bird, Archaeopteryx. ous maniraptorans had combinations of derived He also drew the corollary inference that bipedal­ avian characters, including feathers, teeth without ism preceded the of flight in birds. He serrations, double-condyled quadrates, uncinate

OM66_Text.indd 2 3/31/09 5:54:44 PM origin of birds 3

processes on the ribs, laterally facing glenoid sock­ 2003; Hwang et al. 2002, fig. 31), has challenged ets, retroverted pubes, and (Weisham­ the inference that originated in curso­ pel et al. 2004, and references therein; Kurochkin rial taxa. Additional material indicates that Mi- 2006a, b). The BMT hypothesis has been supported croraptor was probably arboreal and that it was by a large number of cladistic and other studies at least able to glide and was probably capable of (e.g., Padian and Chiappe 1998, Chiappe and Dyke powered flight involving all four limbs (Xu et al. 2002, Clark et al. 2002, Paul 2002, Witmer 2002, 2003, 2005; Martin 2004; Feduccia et al. 2005, 2007; Padian 2004, Weishampel et al. 2004, Xu 2006, and Burnham 2007). Indeed, is strikingly references therein). Some matrices now include similar to the tetrapteryx (four-winged) stage in 100 taxa or more and several hundred characters avian evolution proposed by Beebe (1915, fig. 1; (e.g., Sereno 1999, Holtz et al. 2004). Xu et al. 2003). Such data suggest that the arbo­ Recent books (Dingus and Rowe 1998, Gau­ real model for the origin of bird flight should be thier and Gall 2001, Chiappe and Witmer 2002, reconsidered. Among BMT proponents, Burgers Currie et al. 2004, Weishampel et al. 2004, Norell and Chiappe (1999), Burgers and Padian (2001), and Ellison 2005, Chiappe 2007) summarize these and Padian (2001a, 2004) have retained the cur­ data and assert that the BMT hypothesis has sorial rationale, but Chatterjee (1997), Witmer revolutionized our view of all Archosauria. The (2002), Zhou (2004), and Chatterjee and Templin hypothesis has even been hailed as one of the (2004) have contended that an aboreal origin of major discoveries of 20th-century and a flight can be incorporated into the BMT hypoth­ confirmation of the validity of modern system­ esis. This disagreement among BMT advocates atics (Padian 2001b, 2004; Prum 2002, 2003). As has not led them to question the basic tenets of new data have become available, they have been their hypothesis. viewed as fully supportive (Padian and Chiappe A major concern of critics of the BMT hypoth­ 1998; Chiappe and Dyke 2002; Prum 2002, 2003; esis has been with the assumptions of homology Xu et al. 2003; Xu 2006), and, except for a few ca­ incorporated into matrices supporting it (e.g., veats (Witmer 2002), further debate has been dis­ Martin 1991; Tarsitano 1991; Feduccia 1999, 2002). couraged (Padian 2001b; Prum 2002, 2003). Prum, Inferences of homologies between birds and for example, has called for the integration of or­ theropods in characters of the carpus, manus, nithology as a subfield of dinosaur and tarsus have been challenged (e.g., by Burke and urged that textbooks be rewritten in accord and Feduccia 1997; Feduccia 1999, 2002; Feduc­ with the BMT hypothesis (Prum 2002, 2003). cia and Nowicki 2002; Martin 2004; Feduccia et In spite of the general consensus in support of al. 2005, 2007). As argued by Czerkas et al. (2002), the BMT hypothesis, concern over some of its ele­ Feduccia (2002), Martin (2004), and Feduccia ments has persisted. Paul (2002) has argued that et al. (2005, 2007), the possible avian status of the reason some maniraptoran taxa possess so some maniraptorans could reopen the question many derived avian apomorphies is that they are, of avian ancestry outside of Dinosauria, despite in fact, secondarily flightless birds that are more Paul’s (2002) insistence to the contrary. derived than avian taxa like Archaeopteryx. In light of these ongoing concerns and the Although Paul (2002) retained a theropod ances­ continuing accumulation of new and surpris­ try for birds, support for his hypothesis would ing data from the record, Zhou (2004) ar­ clearly complicate the consensus BMT view. A gued that abandoning the debate concerning the few cladistic analyses have retrieved Alvarez­ origin of birds was premature. We concur with sauridae (e.g., Perle et al. 1993, 1994; Chiappe this assessment and wish to evaluate the BMT et al. 1998) and Oviraptorosauria (Lü et al. 2002, hypothesis in the context of its alternatives. Our Marya´nska et al. 2002) as birds more derived goals are (1) to assess whether the BMT has been than Archaeopteryx, and other noncladistic stud­ as critically tested as is claimed in most of the ies have proposed avian status for various ovi­ literature and (2) to evaluate alternatives to the raptorosaur (Elzanowski 1999, Lü et al. 2005) and BMT hypothesis within a comparative phyloge­ dromaeosaur taxa (Czerkas et al. 2002, Burnham netic framework, including the possibility that 2007). These studies have provided support for some currently nonavian maniraptorans actually elements of Paul’s (2002) hypothesis. belong within Aves. At a minimum, a simultane­ The discovery of Microraptor, described as a ous evaluation of all hypotheses for the origin small basal dromaeosaurid (e.g., Xu et al. 2000, of birds would allow for the possible refutation

OM66_Text.indd 3 3/31/09 5:54:45 PM 4 ORNITHOLOGICAL MONOGRAPHS NO. 66

Fig. 1. Alternative relationships of birds and selected maniraptorans. (A) The sister group of Aves is within the maniraptoran theropod dinosaurs as in Figure 2 (the BMT, birds-are-maniraptoran-theropod-dinosaurs topol­ ogy). (B) Selected maniraptorans are included within Aves.

of the BMT hypothesis. First, we define Aves as Aves as defined here, rather than the topology of used here. Then we describe alternatives to the Figure 1A. BMT hypothesis, outline and discuss the ana­ lytical methods used here, and present two new The BMT Hypothesis and Alternatives analyses. One analysis will be directed toward the first goal of our study, and the other toward All authors agree that the origin of birds lies the second. among the Archosauria (sensu Benton 1999, 2004; Fig. 2). Archosauria is a clade of diapsid Definition of Aves principally characterized by the presence of the in the skull, the lateral man­ Following Chiappe (1992), Chatterjee (1997), dibular fenestra, and serrated teeth implanted and traditional usage, we define Aves (birds) as in sockets (Benton 1999, 2004). The osteology of a node-based clade that includes Archaeopteryx, basal archosaurian taxa such as Proterosuchus modern birds, their most recent common ancestor, (Cruickshank 1972), (Ewer 1965), and and all its descendants (Fig. 1A and B). The term Erythrosuchus is well understood, though basal “bird” refers to any member of this clade. The term relationships are not entirely clear. The crown “birdlike” describes any taxon whose morphology group of Archosauria is Avesuchia (Benton 1999, is like that of members of the clade Aves. Note that 2004), and it includes the and the Aves as here defined is equivalent to sensu Avemetatarsalia. Sereno (1991) argued that cru­ Gauthier (1986) and other workers (e.g., Perle et rotarsal archosaurs constitute a monophyletic al. 1993, 1994; Norell et al. 2001; Clark et al. 2002; lineage: they share a rotary ankle with modified Marya´nska et al. 2002). Whenever the statement calcaneal tubera and condyles, and they lack a is made that some maniraptorans may have been mesotarsus (an ankle joint with a simple hinge birds, the topology of Figure 1B is implied, in between the astragalus/calcaneum and the which these maniraptoran taxa are within the clade rest of the foot). Ingroup relationships among

OM66_Text.indd 4 3/31/09 5:54:46 PM origin of birds 5

Fig. 2. generally accepted phylogeny of the Archosauria including the topology of the BMT hypothesis. Based on Clark et al. (2002), Benton (2004), Langer (2004), and Feduccia (2006).

the Crurotarsi are acknowledged to be un­ 2003). Before they can be adequately evaluated clear (Gower and Wilkinson 1996; Benton 1999, in a comparative phylogenetic framework, they 2004), but one major clade within Crurotarsi is must be stated explicitly, features of the alterna­ the Crocodylomorpha, which includes extant tive topology must be identified, and alternative alligators, , and gavials. Avemetatar­ sister taxa with which Aves may be aligned must salians are diagnosed by their well-developed be specified. In the outline of alternative hypoth­ mesotarsus and hind- modifications for bi­ eses below, and in the analyses reported here, pedalism, particularly their elongate tibiae and we have taken some of these steps. Five major compact elongate metatarsi with reduced fifth hypotheses presently address the origin of birds metatarsals (Benton 1999). The present con­ (Fig. 3A–E). sensus is that the Avemetatarsalia include the Pterosauria and the Dinosauria, including Aves The BMT Hypothesis (Benton 1999). Unfortunately, alternatives to the BMT hypoth­ The widely accepted BMT hypothesis (Fig. 3A) esis have not been explicitly formulated (e.g., states that the sister group of birds lies among Padian and Chiappe 1998, Padian 2001b, Prum the maniraptoran theropod dinosaurs. As noted

OM66_Text.indd 5 3/31/09 5:54:47 PM 6 ORNITHOLOGICAL MONOGRAPHS NO. 66

Fig. 3. five major hypotheses for the origin of birds. (A) The hypothesis that birds are maniraptoran theropod dinosaurs (BMT) and the sister group of birds is the Deinonychosauria (or the Dromaeosauridae) (e.g., Padian 2004). (B) The neoflightless-theropod hypothesis, that the sister group of birds lies among the coelurosaurian theropod dinosaurs but not among the Maniraptora, which are viewed as flightless birds (Paul 2002). (C) The early-archosaur hypothesis, that the sister group of birds is an early arboreal archosaur like Longisquama (e.g., Martin 2004). (D) The crocodylomorph hypothesis, for which various topologies have been proposed: a sister- group relationship between birds and crocodylomorphs (represented by Alligator, Dibrothosuchus, Terrestrisuchus, and ) (Walker 1972) and Aves as nested within Crocodylomorpha but branching off before Crocodylia, which is represented by Alligator, or with Aves as the sister clade of Crocodylia (Whetstone and Whybrow 1983). The topology shown is a polytomy that could be resolved in favor of any of these topologies. (E) The hypothesis that “birds” evolved twice: one lineage is Archaeopteryx, , and the Maniraptora, and the other is (Kurochkin 2006a, b). Hypotheses B, C, and D include or could include the topology shown in F, that at least three clades of maniraptorans (Dromaeosauridae, Troodontidae, and Oviraptorosauria) were radiations within Aves, whose members were at varying stages of flight loss or flight, as advocated by Czerkas et al. (2002), Feduccia (2002), Paul (2002), Martin (2004), and Feduccia et al. (2005, 2007).

OM66_Text.indd 6 3/31/09 5:54:48 PM origin of birds 7

above, the Maniraptora presently include the structures are preserved (Xu et al. 1999b, Ji et al. clades Oviraptorosauria, Troodontidae, Dromaeo­ 2001, Norell and Makovicky 2004). sauridae, Alvarezsauridae, Therizinosauroidea, The controversial alvarezsaurids were small and Aves. The most birdlike maniraptorans are the cursors with extremely reduced and specialized oviraptorosaurs, dromaeosaurs, and troodontids, forelimbs. They were originally placed within which we will refer to as “core maniraptorans.” Aves as defined here on the basis of characters in Oviraptorosaurs were small to medium-sized the skull, pectoral girdle, and hind limb (Perle et with highly specialized that were al. 1993, 1994; Chiappe et al. 1998). Recent cladis­ often ornamented with crests. These skulls were tic analyses have placed them variously as the sis­ heavily pneumatized and usually toothless. ter group to Aves (Chiappe 2002a, Chiappe et al. Some oviraptorosaurs had pectoral girdles and 2002), in a clade with ornithomimosaurs (Sereno forelimbs like those of birds (e.g., Heyu- 1999, 2001), or basal within the Maniraptora annia; Lü et al. 2005), and at least one had a pygo­ (Norell et al. 2001, Clark et al. 2002, Novas and style (; Barsbold et al. 2000). The basal Pol 2002). Padian (2004) classifies them as basal oviraptorosaur had unambiguously birds. Small fibrous structures associated with the vaned feathers in a modern arrangement (Ji et al. of the alvarezsaurid have been 1998, Feduccia et al. 2005). identified as feathers by some authors (e.g., Sch­ Troodontids were small, gracile predators. The weitzer et al. 1999, Schweitzer 2001, Paul 2002). braincases and otic regions of troodontids were Therizinosauroids were bizarre, ground-sloth- exceptionally birdlike (Currie 1985, Currie and like forms; unlike most other maniraptorans, Zhao 1993b). Their teeth were similar to those of they were not particularly birdlike, though they Mesozoic birds (Currie 1987, Norrell and Hwang had opisthopubic pelves (Clark et al. 2004). Most 2004). They possessed a large semilunate carpal of them had small skulls with coarsely serrated, similar to that of birds (Makovicky and Norell lanceolate teeth; vertebrae similar to those of 2004). Unfortunately, troodontid skeletons are other maniraptorans; and exceptionally long poorly represented in the fossil record. At pres­ manual unguals. The phylogenetic relationships ent, no preserved integument has been found of therizinosauroids have been contentious, and with troodontid skeletons. their affinity with theropods has been questioned Dromaeosaurs were small to large predators (Gauthier 1986, Barsbold and Marya´nska 1990), with birdlike opisthopubic pelves and pectoral but they are, by consensus, considered thero­ girdles with ossified sterna and furculae (Norell pods (Clark et al. 2004), and the discovery of and Makovicky 2004). As noted by Ostrom (1969, basal forms like (Kirkland et al. 2005b) 1975, 1976a, b), dromaeosaurs also possessed a has revealed previously unknown maniraptoran large semilunate carpal similar to that of birds. characteristics. Examples are hypapophyses in The distal caudal vertebrae of dromaeosaurs had the presacral vertebrae and the distal placement characteristic long extensions of the prezygapo­ of the obturator process of the (Kirkland physes, a character also seen in the fully flighted et al. 2005b). The basal therizinosauroid Beipiao- Pterosauria. As stated above, the diminutive saurus (Xu et al. 1999a) was described as possess­ basal dromaeosaur Microraptor was probably ar­ ing simple fibrous structures associated with the boreal and capable of powered flight (Xu et al. skeleton. These have been considered feathers by 2003, 2005; Martin 2004; Feduccia et al. 2005, 2007; some authors (Xu et al. 1999a, Clark et al. 2004). Longrich 2006; Burnham 2007). The skeletons of all three of the best-known basal dromaeosaurs The Neoflightless-theropod Hypothesis (Microraptor, , and ) are more similar to those of basal birds than are The “neoflightless-theropod hypothesis” (Fig. the skeletons of more derived dromaeosaurs like 3B) states that the sister group of Aves is an un­ (Paul 2002; Burnham 2004, 2007). specified and hypothetical lineage of arboreal For the basal status of Bambiraptor, see Burnham coelurosaurian theropod dinosaurs (Paul 2002, G. (2004, 2007) and the cladistic analysis of Senter et Paul pers. comm.). Paul proposes that most and al. (2004). Among specimens of dromaeosaurs, perhaps all of the maniraptoran clades are flight­ only Microraptor possesses unambiguous vaned less or flying lineages within the avian radiation, feathers (Xu et al. 2003; Feduccia et al. 2005, generally according to the topology of Figures 1B 2007; Longrich 2006); in other taxa, only fibrous and 3F, but he is unsure of the boundaries of Aves,

OM66_Text.indd 7 3/31/09 5:54:48 PM 8 ORNITHOLOGICAL MONOGRAPHS NO. 66

so the sense in which he considers maniraptorans James and J. Pourtless pers. obs.; see Table 1) sup­ to be flying or flightless birds is not entirely clear ports the classification of Sharov (1970) and Jones (Paul 2002, G. Paul pers. comm.). To formulate et al. (2000, 2001). Martin (2004) elaborated on the Paul’s hypothesis explicitly, we have interpreted osteological similarities between Longisquama and his position as generally congruent with the to­ birds in dentition, characters of the skull, and the pology of Figures 1B and 3F. His hypothesis is a presence of a boomerang-shaped similar modification of the BMT hypothesis, in that, al­ to that of basal birds. Unfortunately, the pelvic though he places the maniraptorans within Aves, girdle and hind limb are not known. Longisquama he still places both within the Theropoda. Paul is best considered a basal archosaur of uncertain (2002) summarized evidence for this hypothesis affinity (seeF ig. 2). Although the type specimen of and presented a tabular, but not a cladistic, analy­ Longisquama is incomplete, it has been described sis of the avian features of maniraptorans. in greater detail than relevant alternatives, and it is the only “early archosaur” that has been explic­ The Early-archosaur Hypothesis itly connected with the origin of birds by defend­ ers of the early-archosaur hypothesis. To date, The “early-archosaur hypothesis” (Fig. 3C) no version of the early-archosaur hypothesis has states that the origin of birds is more likely to be been evaluated by means of cladistics. among early archosaurs than among the theropod dinosaurs (e.g., Tarsitano and Hecht 1980; Feduccia The Crocodylomorph Hypothesis and Wild 1993; Welman 1995; Feduccia 1999, 2002; Czerkas and Yuan 2002; Czerkas et al. 2002; Martin The “crocodylomorph hypothesis” (Fig. 3D), 2004; Feduccia et al. 2005, 2007). As presently un­ first proposed by Walker, states that birds share derstood, this hypothesis includes the propositions an immediate common ancestor with Crocodylo­ that most maniraptorans are flying and flightless morpha (Walker 1972, 1977, 1990), or that the sister lineages within Aves (as in Figs. 1B and 3F) and group of birds is within the Crocodylomorpha but that they are, in fact, not theropod dinosaurs (Czer­ outside of the Crocodylia, or that Aves is the sister kas et al. 2002; Feduccia 2002; Martin 2004; Feduc­ clade of Crocodylia (Martin et al. 1980, Whetstone cia et al. 2005, 2007). According to this alternative, and Whybrow 1983, Martin and Stewart 1999). the Theropoda as presently constituted are not Any of these versions of the crocodylomorph hy­ monophyletic. Aves, including various manirap­ pothesis can be modified to include the proposition torans, is not nested inside Theropoda. Similarities that some or all maniraptoran clades belong within between nonmaniraptoran theropods and birds are Aves (as in the neoflightless-theropod and early- accounted for by homoplasy. archosaur hypotheses and as depicted in Figs. 1B Temporally early and phylogenetically basal ar­ and 3F), following the current understanding of chosaurs are not a monophyletic group, so a rep­ the early-archosaur hypothesis. We have modified resentative lineage or taxon must be designated the crocodylomorph hypothesis accordingly. To from among basal archosaurs with which Aves date, no version of the crocodylomorph hypothesis may be aligned. Some candidates, like Euparkeria has been evaluated by means of cladistics. (Broom 1913, Welman 1995), are no longer under The crocodylomorph hypothesis has not re­ consideration (see Gower and Weber 1998). The ceived much attention in the literature. Walker best-studied current candidate for a potential ar­ (1985) recanted it and then later supported it again chosaurian ancestor or sister taxon is Longisquama. (Walker 1990). Note that crurotarsal archosaurs Sharov (1970) noted similarities to birds in the underwent an extensive adaptive radiation in the skeleton and integument of Longisquama and sug­ , and they were far more diverse in the gested that it may be close to avian ancestry. Jones early Mesozoic than they are today (Benton 2004). et al. (2000, 2001) described the birdlike osteologi­ For example, crocodylomorphs like Terrestrisuchus cal characters and the featherlike morphology of were gracile (Crush 1984), and other noncroco­ the integumentary appendages of Longisquama, dylomorph crurotarsal taxa, like and but the latter was disputed by Prum et al. (2001) , were medium-sized to large forms and Unwin and Benton (2001). Unwin and Ben­ convergent on theropod dinosaurs. Taxa like Pos- ton (2001) and Senter (2004) questioned the status tosuchus, Ornithosuchus, and were at least of Longisquama as an archosaur, but its antorbital facultatively bipedal and had many pelvic-girdle, fenestra (Jones et al. 2000, 2001; Martin 2004; F. hind-limb, and cranial characters similar to those

OM66_Text.indd 8 3/31/09 5:54:49 PM origin of birds 9

Table 1. Taxa included in our analysis and the principal references used for coding their characters.

Taxon Group References

Alligator Crocodylomorpha Owen (1850), Reese (1915), Mook (1921), Iordansky (1973), personal observation of skulls and postcrania in the collection of the Florida State University Department of Biological Science Neotetanurae Madsen (1993), Chure (2001), Clark et al. (2002), Holtz et al. (2004) Therizinosauroidea Russell and Dong (1993b), Clark et al. (2002, 2004) Ornithurae Clarke and Norell (2002) Archaeopteryx a Aves Walker (1980, 1985), Martin (1985, 1991, 1995, 2004), Elzanowski and Wellnhofer (1996), Elzanowski (2002), Mayr et al. (2005, 2007), personal observation of casts in the collection of the University of Museum of Natural History (UKMNH) Oviraptorosauria Kurzanov (1982, 1983, 1985), Norman (1990), Clark et al. (2002), Marya´nska et al. (2002), Vickers-Rich et al. (2002), Osmólska et al. (2004) Bambiraptor Dromaeosauridae Burnham et al. (2000), Burnham (2004), Norell and Makovicky (2004), personal observation of type material and casts at UKMNH Ornithurae Martin and Tate (1976), Martin and Bonner (1977), Galton and Martin (2002) Troodontidae Norell et al. (2000), Clark et al. (2002), Makovicky et al. (2003), Makovicky and Norell (2004) Caudipteryx Oviraptorosauria Ji et al. (1998), Zhou et al. (2000), Padian et al. (2001), Clark et al. (2002), Marya´nska et al. (2002), Paul (2002), Osmólska et al. (2004) Madsen and Welles (2000), Tykoski and Rowe (2004) Oviraptorosauria Clark et al. (1999, 2001, 2002), Osmólska et al. (2004) Coelurosauria Carpenter et al. (2005) Compsognathusb Coelurosauria Ostrom (1978) Oviraptorosauria Clark et al. (2002), Osmólska et al. (2004) Aves Martin et al. (1998b), Chiappe et al. (1999), Hou et al. (1999b), Zhou and Hou (2002) Deinonychus Dromaeosauridae Ostrom (1969, 1974), Clark et al. (2002), Norell and Makovicky (2004) Dibrothosuchus Crocodylomorpha Wu and Chatterjee (1993) Dilong Xu et al. (2004) Ceratosauria Welles (1984), Tykoski and Rowe (2004) Dromaeosauridae Currie (1995), Clark et al. (2002), Norell and Makovicky (2004) Effigiac Crurotarsi Nesbitt and Norell (2006) Ornithurae Elzanowski and Galton (1991), Galton and Martin (2002) Enantiornithes Hou et al. (1999a), Chiappe and Walker (2002) Saurischia Langer (2004) Therizinosauroidea Clark et al. (1994, 2002, 2004) Erpetosuchus Crurotarsi Benton and Walker (2002) Erythrosuchus Archosauria Gower (1997, 2001, 2003) Euparkeria Archosauria Ewer (1965), Gower and Weber (1998) Falcariusd Therizinosauroidea Kirkland et al. (2005b) Osmólska et al. (1972), Barsbold and Osmólska (1990), Hurum (2001), Clark et al. (2002), Makovicky et al. (2004) Ornithurae You et al. (2006) Tyrannosauroidea Xu et al. (2006) Ornithomimosauria Makovicky et al. (2004), Kobayashi and Barsbold (2005) Herrerasauruse Saurischia Novas (1993), Sereno (1993), Sereno and Novas (1993), Langer (2004) Crocodylomorpha Colbert (1952), Clark et al. (2000) Oviraptorosauria Lü (2002), Lü et al. (2005) Ornithurae Zhou and Zhang (2005) Huaxiagnathus Hwang et al. (2004) Enantiornithes Sanz and Bonaparte (1992), Sanz et al. (2002) Ornithurae Martin and Stewart (1977), Chiappe (2002a), Clarke (2004) Oviraptorosauria Xu et al. (2002a), Osmólska et al. (2004), Senter et al. (2004). Senter et al. (2004) argued that Incisivosaurus is a junior of , but this argument is not accepted here, and the two taxa are considered distinct.

(continued)

OM66_Text.indd 9 3/31/09 5:54:49 PM 10 ORNITHOLOGICAL MONOGRAPHS NO. 66

Table 1. (Continued)

Taxon Group References

Ingenia Oviraptorosauria Clark et al. (2002), Lü et al. (2002), Osmólska et al. (2004) Aves Zhou and Zhang (2002, 2003b) Compsognathidae Göhlich and Chiappe (2006) Longisquama Archosauria Sharov (1970), Jones et al. (2000, 2001), Martin (2004), personal observation of unpublished photographs of type and casts at UKMNH Sereno and Arcucci (1994) Microraptor Dromaeosauridae Xu et al. (2000, 2003), Clark et al. (2002), Hwang et al. (2002), Norell and Makovicky (2004) Oviraptorosauria Makovicky and Sues (1998), Clark et al. (2002), Osmólska et al. (2004) Alvarezsauridae Perle et al. (1993, 1994), Clark et al. (2002) Therizinosauroidea Kirkland and Wolfe (2001), Clark et al. (2004), Kirkland et al. (2005a) Carpenter et al. (2005) Ornithosuchus Crurotarsi Walker (1964) Oviraptorosauria Clark et al. (2002), Osmólska et al. (2004) Ornithurae Chiappe (2002a, b), personal observation of cast of MACN-N-11 at UKMNH Pelecanimimus Ornithomimosauria Pérez-Moreno et al. (1994), Clark et al. (2002), Makovicky et al. (2004) Postosuchus Crurotarsi Chatterjee (1985) Proterosuchus Archosauria Haughton (1924), Cruickshank (1972), Clark et al. (1993) Enantiornithes Zhang and Zhou (2000) Incertae sedis Forster et al. (1998), Clark et al. (2002) Aves Zhou and Zhang (2003a) Troodontidae Clark et al. (2002), Makovicky and Norell (2004). We treated the as a composite taxon. Avemetatarsalia Benton (1999) Shuvuuia Alvarezsauridae Chiappe et al. (2002), Clark et al. (2002), Suzuki et al. (2002) Enantiornithes Martin and Zhou (1997), Sereno et al. (2002), supplemented by personal observation of casts of BVP 538a and IVPPV 9769 at UKMNH. Our analysis follows the suggestion of Sereno et al. (2002) that Sinornis and are synonymous. Troodontidae Russell and Dong (1993a), Currie and Dong (2001), Clark et al. (2002), Makovicky and Norell (2004) Ornithomimosauria Kobayashi and Lü (2003), Makovicky et al. (2004) Sinornithosaurus Dromaeosauridae Xu et al. (1999b), Ji et al. (2001), Xu and Wu (2001), Norell and Makovicky (2004). NGMC 91 (Ji et al. 2001) is here considered to be a referred specimen of Sinornithosaurus. Sinosauropteryx Compsognathidae Currie and Chen (2001) Incertae sedis Xu et al. (2002b), Makovicky and Norell (2004) Neotetanurae Currie and Zhao (1993a), Holtz et al. (2004) Sphenosuchus Crocodylomorpha Walker (1972, 1990), unpublished data and private correspondence, provided courtesy of A. Feduccia Syntarsus Ceratosauria Raath (1985), Colbert (1989), Rowe and Gauthier (1990), Tykoski and Rowe (2004). We concur with Paul (2002) in regarding Syntarsus and as nearly identical and therefore probably synonymous. Terrestrisuchus Crocodylomorpha Crush (1984) Troodontidae Currie (1985, 1987), Currie and Zhao (1993b), Clark et al. (2002), Makovicky and Norell (2004) Tyrannosauroidea Clark et al. (2002), Brochu (2003), Holtz (2004) Dromaeosauridae Novas and Puerta (1997) Velociraptor Dromaeosauridae Norell and Makovicky (1997, 1999, 2004), Barsbold and Osmólska (1999) Ornithurae Zhou and Zhang (2001)

a We regard Archaeopteryx to be monospecific, followingH ouck et al. (1990), Senter and Robins (2003), and Bennett (2008), as opposed to Elzanowski (2002), Mayr et al. (2005, 2007), and Christiansen (2006). b Peyer (2006) came to our attention too late for us to refer to it in scoring . c Nesbitt (2007) came to our attention too late for us to refer to it in scoring Effigia. d Zanno (2006) came to our attention too late for us to refer to it in scoring Falcarius. e Sereno (2007) came to our attention too late for us to refer to it in scoring .

OM66_Text.indd 10 3/31/09 5:54:50 PM origin of birds 11

of dinosaurs (Romer 1956, Walker 1964, Chatterjee test phylogenetic hypotheses but as an exploratory 1985, Feduccia 1999, Nesbitt 2007, and references method, useful, if handled sensitively, for compar­ therein). In all cases in which crurotarsal archo­ ing and evaluating hypotheses. saurs developed , their pelvic girdles and hind-limb morphology converged on those of Methods avemetatarsalian archosaurs (Romer 1956, Walker 1964, Chatterjee 1985, Feduccia 1999, Nesbitt 2007, Construction of Matrices and references therein). An excellent example is provided by the morphology of the crurotar­ In pursuing our dual goals of (1) evaluating sal Effigia, which is especially convergent on the the extent to which the BMT hypothesis has been morphology of ornithomimosaurs and shows fur­ tested and (2) evaluating the BMT and its alterna­ ther convergence on numerous characters of the tives in a single phylogenetic framework, we con­ avemetatarsalian, dinosaurian, theropod, neoteta­ sidered two sets of data. The first is a generally nurine, and coeluosaurian skeletons (Nesbitt and accepted matrix from a recently published paper Norell 2006, Nesbitt 2007). (Clark et al. 2002) that is representative of the on­ going research of the Theropod Working Group The Hypothesis that Birds Evolved Twice (see Acknowledgments). In this version, which we call the “CNM matrix” (for Clark-Norell-Ma­ The hypothesis of Kurochkin (2006a, b) that kovicky), the taxa are 43 theropod dinosaurs and “birds” evolved twice (Fig. 3E) states that a lineage three birds, scored for 208 characters. The intent that includes Archaeopteryx and Enantiornithes of Clark et al.’s (2002) study was to examine re­ is nested within Maniraptora, and Maniraptora lationships within the ingroup, which consisted is nested within Theropoda, but the Ornithurae of birds and coelurosaurian theropods. In the are aligned with another older, nondinosaurian CNM matrix, 52% of cells are scored as not ap­ archosauromorph lineage. Kurochkin’s Archo­ plicable or missing data. The matrix includes 40 sauromorpha are equivalent to Archosauria as multistate characters, 13 of which are ordered. defined here (E. Kurochkin pers. comm.). Kuro­ We digitized the published matrix to reanalyze it. chkin’s hypothesis has not been evaluated by Evaluation and reanalysis of this typical sample means of cladistics. matrix allows evaluation of the support such ma­ trices offer the BMT hypothesis. Cladistics To evaluate the BMT hypothesis simultane­ ously with its alternatives, we reviewed the lit­ Our cladistic approach follows the recent recom­ erature in detail, examined specimens (Appendix mendations of Jenner (2004), who emphasized that 1), and evaluated potential homologies of charac­ great care must be taken in analyzing the compara­ ters across taxa in a broad array of archosaurs. We tive morphology of taxa to minimize subjectivity then constructed a new 79-taxon × 221-character and bias. Jenner insists that all potentially relevant matrix. An additional 21 characters were included and usable corroborating and disconfirming evi­ in an alternative analysis (see below). With the dence for all hypotheses under consideration be additional 21 characters, our new matrix contains included in the matrix. Taxon selection should in­ 242 characters. Appendix 2 gives the character clude all currently proposed sister groups, and the list, notes on some of the characters, and their matrix should be conservative in the selection and sources; Appendix 3 discusses our reasons for scoring of characters. A conscious effort should be excluding some important characters; Appendix made to minimize a-priori assumptions of homol­ 4 gives the matrix; and Appendix 5 gives com­ ogy, particularly where the anatomical data are ments on scores for certain taxa. To understand ambiguous. The logical integrity of an analysis our study fully, please consult the appendices. would be compromised if statements of homology We included potential sister taxa identified by presupposing particular phylogenetic hypotheses the hypotheses under test. Poorly known taxa were incorporated into the matrix. Jenner (2004) were included if they have figured prominently also insisted that the relationship between the evi­ in discussions of the origin of birds or the phylog­ dence (the matrix) and the optimal tree(s) should eny of birds or if they had to be included because be examined with a routine like the bootstrap pro­ of the limited number of representatives for a par­ cedure. Cladistics should be treated not as a way to ticularly important group. Our taxon list consists

OM66_Text.indd 11 3/31/09 5:54:50 PM 12 ORNITHOLOGICAL MONOGRAPHS NO. 66

of four basal archosaurs, nine crurotarsal archo­ (1972, 1990) and Whetstone and Martin (1979) saurs, two nondinosaurian avemetatarsalians, identified as possibly supporting a relationship two basal nontheropod saurischian dinosaurs, 46 between birds and crocodylomorphs (e.g., our theropod dinosaurs (of which 27 were nonavian character 71). We were, however, as critical with maniraptorans and 19 were nonmaniraptorans), respect to these characters as we were with respect and 16 Mesozoic birds (Table 1). to those in the CNM matrix. For example, many of Of the 221 characters turned on for the primary the cranial characters identified by Walker (1972, analysis of the matrix, 61 are multistate. Twenty 1990) as supporting a close relationship between of the multistate characters are ordered in all our birds and crocodylomorphs were omitted for the analyses; included are 5 characters of the integu­ same reasons that characters from the CNM ma­ ment, 108 characters of the skull, 33 characters trix were omitted. of the axial skeleton, and 75 characters of the ap­ Although quantitative characters are impor­ pendicular skeleton (Appendix 2). Our character tant in phylogenetic analysis, obtaining homolo­ states are not polarized (i.e., the zero state is not gous measurements across the broad span of assumed a priori to be the most primitive state taxa in our matrix was difficult, so the number for the character). For selection of characters, we of quantitative characters had to be minimized. used the CNM data matrix as a template. Our ma­ Some important characters of birds that could trix shares with the CNM matrix many characters not be excluded (e.g., our character 170) may best common to other phylogenetic analyses involving be expressed quantitatively, but we used them as theropod or other archosaurian taxa. We retained qualitative characters. For treatment of some es­ 96 of their 208 characters verbatim. We excluded pecially controversial characters, like the metotic characters that had continuous variation (e.g., fissure, the dentition, the , and the their character 110), characters that were propor­ furcula, see the notes accompanying Appendix 2. tions or that should be formulated as proportions Following the recommendations of Rieppel (e.g., their characters 88 and 142), and characters and Kearney (2002), Jenner (2004), and Kearney for which we could not understand or detect and Rieppel (2006) in constructing our matrix, we the variation intended by the authors (e.g., their tried to minimize making a-priori assumptions character 120). Norell et al. (2001) and Clark et al. about the homology of characters across taxa. Pri­ (2002) used similar criteria in determining which mary homology statements that are not indepen­ characters to use in their matrices, and both noted dently testable, when used as characters and later that many characters used in the literature are un­ regarded as corroborated statements of secondary satisfactory. Some of the characters in the CNM homology (i.e., synapomorphies; de Pinna 1991), matrix were designed to address specific ques­ can be misleading. If the data are ambiguous, they tions in coelurosaur phylogeny that were not lead to unjustified confidence in the phylogeny. If relevant to our analysis (e.g., their character 156), the primary homology statements assume a fa­ so they were excluded. We reformulated some vored hypothesis, they threaten the analysis with characters (e.g., their character 148). Our criteria circularity (Rieppel and Kearney 2002, Kearney for excluding characters from the CNM matrix and Rieppel 2006). Our objective was to create a largely match those of Livezey and Zusi (2006). matrix that, to the extent possible, incorporated To be able to evaluate the BMT against alterna­ characters for which homologies are best sup­ tive hypotheses, we added, in addition to the thero­ ported by the evidence and excluded characters pod characters from the CNM matrix, characters for which homologies rely on the implications of from other published analyses that are potential specific hypotheses. Unfortunately, to accommo­ synapomorphies of various nondinosaurian ar­ date the recommendations of Rieppel and Kearney chosaur clades and characters that have been pro­ (2002) and Kearney and Rieppel (2006), we had posed as uniting birds with other archosaur taxa to exclude several major structures from the pri­ (see references accompanying the character list in mary analysis of our matrix. They were included Appendix 2). For example, we included dental in our matrix but were turned on only for the al­ characters that Martin et al. (1980), Martin (1983, ternative analysis; see below. Excluded from the 1991), and Martin and Stewart (1999) suggested primary analysis of our matrix were elements of might support a close relationship between birds the palate, the basipterygoid process, the carpus, and crocodilians (e.g., our characters 91–95) and the manus, and the tarsus, together accounting for braincase and otic-region characters that Walker the exclusion of 18 characters (characters 12–15,

OM66_Text.indd 12 3/31/09 5:54:50 PM origin of birds 13

Fig. 4. embryonic chicken wing bud at (A) 5.5 days of development and (B) 7.5 days of development. Note that in B the regression of the ulnare (indicated by a dashed line) is followed by enlargement of the pisiform, which in later stages of development forms the structure usually, but inaccurately, referred to as the “ulnare” in the adult carpus (e.g., Hinchliffe 1985, Baumel and Witmer 1993; see Kundrát 2008 for an alternative view, described in Appendix 3). Abbreviations: dc = distal carpal, R = radius, U = ulna, u = ulnare, p = pisiform, and r = radiale. Modified from figure 21B, C of Feduccia et al. (2005).

63–65, 148–156, and 196–197) used in the CNM and theropods, in spite of the presence of similar matrix. These ambiguities are summarized in the carpals in some maniraptorans. Note that lunate next paragraph. shape is not sufficient to establish homology (see The most recent analysis of the avian palate Fig. 5 and Appendix 3). Because the patterns of (Livezey and Zusi 2006) concluded that homolo­ carpal evolution in theropods are not well under­ gies with theropods and other archosaurs are un­ stood, comparisons between birds and theropods certain (see also McDowell 1978). Disagreement and between theropods and other archosaurs are arises in the literature on the morphology of the difficult (Appendix 3). Moreover, assessment of basal avian palate, particularly the morphol­ carpal homologies is largely contingent on resolu­ ogy of the pterygoid and palatine (compare El­ tion of the identities of the manual digits of birds zanowski and Wellnhofer [1996] and Elzanowski and theropods, which remain contentious (Burke [2002] with Mayr et al. [2005, 2007]). Elzanowski and Feduccia 1997; Feduccia 1999; Wagner and (2002) regarded the palate of Archaeopteryx as Gauthier 1999; Feduccia and Nowicki 2002; Kun­ too autapomorphic to permit clear comparisons drát et al. 2002; Larsson and Wagner 2002; Galis et with other archosaurs, including other Mesozoic al. 2003, 2005; Vargas and Fallon 2005a, b; Welten birds. A structure homologous with the true ba­ et al. 2005; see Appendix 3). According to Wag­ sipterygoid process of reptiles is apparently ab­ ner and Gauthier (1999), manual characters are sent in both modern birds and crocodylomorphs, not necessary for retrieving the consensus BMT though the underlying cartilages from which a topology, which implies that these characters can true basipterygoid process would develop are be excluded without prejudicing an analysis. The present, which complicates assessments of ho­ homology of the ascending sheet of bracing mology (McDowell 1978, Walker 1990). Carpal the distal tibia in birds and theropods is unclear homologies among birds, theropods, and other (Martin et al. 1980; Martin 1983, 1991; Martin and archosaurs are unclear (Hinchliffe and Hecht Stewart 1985; Feduccia 1999). The morphology of 1984, Hinchliffe 1985, Feduccia 1999; Appendix 3). this structure is disputed (compare Martin et al. Autapomorphic features in the development of [1980] and Martin [1991] with Mayr et al. [2005, the neornithine carpus, such as the disappearance 2007]), and the significance of differences in mor­ of the ulnare and its replacement by the pisiform phology and development of these structures in (Fig. 4), complicate comparisons with other archo­ birds and theropods (see Fig. 6) is unclear (see saurs, and theropods in particular (Appendix 3); Appendix 3 for further details). uncertainties in the identity of the elements of the Because of the above ambiguities, these five avian semilunate carpal complicate assessments sets of characters cannot be coded for birds and of homology within Aves, and also between birds theropods without unjustified assumptions of

OM66_Text.indd 13 3/31/09 5:54:51 PM 14 ORNITHOLOGICAL MONOGRAPHS NO. 66

Fig. 5. Carpal cartilages in the chicken wing bud at (A) 9.5, (B) 11.5, and (C) 12.5 days of development. Note the coalescence of the semilunate carpal from the fusion of a distal carpal and element X. The pisiform by 12.5 days of development begins to take on the notched form of the supposed “ulnare” of the adult carpus (see Kundrát 2008 for an alternative view, described in Appendix 3). Abbreviations: mc = metacarpal; all others as in Figure 4. Modified from figures 9–10 of Hinchliffe (1985).

homology. They were not included in the pri­ History and the Royal Tyrrell Museum. These mary analysis of our matrix. This decision is specimens included casts and original material of understood to be especially controversial, so theropods and Mesozoic birds (Appendix 1). In we have documented our reasoning, which was spite of some of the common drawbacks of most based on careful review of the anatomical evi­ cladistic analyses, such as counting the absence of dence, in Appendix 3. Nevertheless, to determine a structure as a character state (Jenner 2004), we whether inclusion of these characters would al­ were not able to avoid the binary coding of the ter our results, we ran an alternative analysis to presence and absence of many characters. With evaluate differences made by their inclusion. A the exclusion of the 21 characters turned on in the total of 21 characters were turned on for the alter­ alternative analysis, 39% of the cells in our matrix native analysis: 1 character for the basipterygoid were coded as missing data, and 3% as not ap­ process, 5 characters of the palate, 14 characters plicable. of the carpus and manus, and 1 character of the tarsus (Appendix 2). When these 21 characters Methods of Analysis of the Matrices were turned on for the alternative analysis, the data matrix contained 242 characters as opposed We used PAUP*, version 4.0b10 (Swofford to 221. Where homologies were uncertain, the 2003), to conduct heuristic searches with both added characters were coded as unknown (signi­ matrices. Heuristic searches for most-parsimo­ fied by a question mark) for those taxa. nious trees (MPTs) were performed by stepwise As part of our effort to be conservative, we addition followed by tree bisection–reconnection scored the controversial integumentary append­ (TBR) branch rearrangement, with a maximum ages of Longisquama as unknown for character 1, tree value of 10,000, as in most of the literature. To the presence of integumentary structures homol­ select trees for interpretation from among a set of ogous with avian feathers. Overall, we attempted MPTs, we used a combination of bootstrapping, to construct a matrix that satisfies, to the extent successive pruning of selected taxa, and 50% possible, the crucial assumption of cladistics, that majority-rule summary trees. For the bootstrap­ the scoring of each character can be viewed as an ping analyses conducted during analysis of both independently testable homology statement (e.g., matrices, 500 pseudoreplicates were generated by Rieppel and Kearney 2002, Kearney and Rieppel fast stepwise addition; groups with a frequency 2006). of >50% were displayed on 50% majority-rule For coding taxa, we relied on the literature summaries of the bootstrapped trees. Such and examination of specimens in the collections methods estimate the best-supported elements of the University of Kansas Museum of Natural of a tree. This procedure is in agreement with the

OM66_Text.indd 14 3/31/09 5:54:52 PM origin of birds 15

Fig. 6. (A) Distal and proximal tarsals of Albertosaurus, (B) a composite of the London and Berlin specimens of Archaeopteryx, and (C) the Thermopolis specimen of Archaeopteryx. The shaded area represents the “ascending process of the astragalus,” the homology of which is unclear across these taxa; in birds it may not be an ascend­ ing process of the astragalus at all, but rather a descending pretibial ossification (see Appendix 3). All tarsi are shown in anterior view and are not to scale. Left tarsi are shown in A and B; C depicts the right tarsus reversed to match the orientation of A and B. In C the are drawn slightly separated for ease of identification. A is modified from figure 10 of Welles and Long (1974), B is modified from figure 1g of Martin et al. (1980), and C is modified from figure 12b of Mayr et al. (2007).

systematic practices advocated by Wilkinson Pruning can be an important step for the ex­ (1996), Wilkinson and Thorley (2001), Sanderson ploratory analysis of majority-rule trees (Wilkin­ and Shaffer (2002), Cranston and Rannala (2007), son 1996, 2003; Steel and Penny 2000; Cranston and Holder et al. (2008). We considered reporting and Rannala 2007; M. Holder pers. comm.; D. a strict consensus of MPTs and adding bootstrap Swofford pers. comm.). It can correct for the values to it, but we found that our method was negative effects of unstable taxa on tree resolu­ able to reveal more statistically supported struc­ tion, especially those taxa with large percentages ture in subsets of the data that addressed particu­ of missing data. It does not delete taxa from the lar questions of interest. matrix and does not involve reanalysis; taxa are

OM66_Text.indd 15 3/31/09 5:54:53 PM 16 ORNITHOLOGICAL MONOGRAPHS NO. 66

pruned from the bootstrapped trees and boot­ the differences. The statistical significance of strap values are recalculated at each pruning to Kishino-Hasegawa tests is reported only to indi­ reflect shifts in support when specified taxa are cate the relative importance of differences among temporarily ignored, and the results are sum­ comparisons. Such a-posteriori tests are biased to­ marized with 50% majority-rule consensus trees. ward finding statistical significance, but some of This procedure permits evaluation of the sensi­ the bias can be controlled if the test is considered tivity of the consensus topology. to be one-tailed rather than two-tailed, at least Pruning was used in both our reanalysis of the when two-tailed P values are >0.1 (Goldman et CNM matrix and the analysis of our new matrix. al. 2000). Any remaining bias would still be in the We agree with Wilkinson (1996) and Cranston direction of finding statistical significance. and Rannala (2007) that, by focusing attention on relationships among exemplar taxa of the groups Results under consideration, pruning can reveal struc­ ture in a tree that would have been obscured by Reanalysis of the CNM Matrix calculation of bootstrap support when all the taxa are taken into consideration. Study of statistical Our reanalysis of the 46-taxon × 208-character bootstrap support for specific branches on trees CNM matrix using PAUP* allowed the same can provide useful information, even when the set of ordered characters as Clark et al.’s (2002). support is low. We initially pruned taxa by per­ Our heuristic search recovered 356 MPTs of 620 centages of missing data, a procedure similar to steps. Following Clark et al. (2002) for their strict the automatic pruning procedures described in consensus tree, we rooted our 50% majority-rule the literature by Wilkinson (1996) and Steel and summary tree of bootstrap replicates using the Penny (2000), but we found that this approach theropods Sinraptor and Allosaurus. In this un­ was insensitive to the potentially useful informa­ pruned tree, birds, maniraptorans, ornithomi­ tion present in some taxa with a high percentage mosaurs, and other coelurosaurs, except for of missing data (see also Kearney 2002, Kearney Tyrannosaurus and Albertosaurus, appeared in a and Clark 2003). Taking the suggestion of D. polytomy with only a few resolved clades (Fig. 7). Swofford (pers. comm.), we then modified our When the tree was pruned to 24 taxa that rep­ approach to pruning, basing it on our ability to resented all the major ingroup clades, however, make predictions and ask specific questions. For bootstrap values showed substantial support for example, if the BMT hypothesis is robust, then the Coelurosauria (75% of bootstrapped trees), when maniraptorans are pruned from the tree, Maniraptoriformes (95%), and a clade of birds Aves should still be nested within a monophyl­ and maniraptorans (75%) in which internal re­ etic Theropoda. lationships were largely unresolved (Fig. 8). The We ran both the CNM matrix and our new ma­ Ornithomimosauria (represented by Struthiomi- trix with various backbone constraints. This pro­ mus and Gallimimus) appeared outside a clade cedure follows recent efforts to place phylogenetic of birds and maniraptorans, as expected. Aves, inference in a more statistical framework (Felsen­ Dromaeosauridae, Deinonychosauria, and Ther­ stein 2001, 2004). See Corfe and Butler (2006) for a izinosauroidea did not appear as clades on the similar application. Because some of the alterna­ pruned 50% majority-rule tree, though the Ovi­ tives to the BMT hypothesis include the propo­ raptorosauria (72%), Troodontidae (73%), and Al­ sition that at least some maniraptorans belong varezsauridae (100%) were well supported. With within Aves, we tried constraining maniraptoran further pruning to 18 taxa, Aves and Dromaeo­ taxa to be within Aves. Differences in lengths sauridae were resolved as two branches of a poly­ of MPTs were obtained from heuristic searches tomy including troodontids and oviraptorosaurs when topological constraints were enforced, and (not shown). Kishino-Hasegawa tests (Kishino and Hasegawa When we used backbone constraints to con­ 1989) were used to evaluate these differences. In strain various maniraptoran groups and com­ the analysis of our new matrix, results from the binations of maniraptoran groups to be within application of backbone constraints correspond­ Aves as defined here, MPTs were only 0–5 steps ing to the topologies of the alternative hypotheses longer than unconstrained MPTs, increasing from of the origin of birds (Fig. 3) were compared, and the alvarezsaurids (no difference, equally parsi­ Kishino-Hasegawa tests were used to evaluate monious), to the troodontids and dromaeosaurs

OM66_Text.indd 16 3/31/09 5:54:54 PM origin of birds 17

Fig. 7. A 50% majority-rule summary of 500 bootstrap replicates of most-parsimonious trees calculated from the matrix reported by Clark et al. (2002). In this tree, birds, maniraptorans, ornithomimosaurs, and other coelu­ rosaurs, except for Tyrannosaurus and Albertosaurus, appear in a polytomy with only a few resolved clades.

(1 step), to the oviraptorosaurs and therizinosau­ constraints. Kishino-Hasegawa tests of differ­ roids (3 steps), to the combination of oviraptoro­ ences in the lengths of unconstrained MPTs and saurs, troodontids, and dromaeosaurs (5 steps) MPTs resulting from the application of the various (Table 2). An increase to 26 steps in the length of constraints are reported as the range of standard MPTs resulted when the ornithomimosaurs were deviations among five pairs of randomly selected so constrained. The numbers of MPTs increased MPTs, their average t value, and the average P substantially (to >10,000) with 7 of the 13 various value for one-tailed tests. The results of the first

OM66_Text.indd 17 3/31/09 5:54:55 PM 18 ORNITHOLOGICAL MONOGRAPHS NO. 66

Fig. 8. As in Figure 7, but pruned to 24 taxa that represent all the major ingroup clades of Clark et al. (2002). Bootstrap values show substantial support for the Coelurosauria (75%), Maniraptoriformes (95%), and a clade of birds and maniraptorans (75%) in which internal relationships are largely unresolved. The polytomy of birds and maniraptorans indicates that the interrelationships of birds and maniraptorans are not resolved by our analysis of this data set.

nine tests were not significant. See the asterisk in troodontids, and dromaeosaurs), we pruned away Table 2 for an explanation of the potential bias in the other maniraptorans (therizinosauroids and al­ this test. varezsaurids) (Fig. 9). With the 20 remaining taxa, Aves was not recovered, but the basal avian taxa Analysis of the New Matrix (Archaeopteryx, Confuciusornis, Sapeornis) were in a weakly supported clade (58%) with the manirap­ When we analyzed the new 79-taxon × 221- torans and Longisquama. Interrelationships within character matrix without topological constraints this clade were largely unresolved. This clade was using PAUP*, we recovered 360 MPTs of 1,355 in a polytomy with a weakly supported clade of steps. We used the basal archosaur Proterosuchus nonmaniraptoriform theropods (66%) and Cro­ for rooting. As with the analysis of the CNM ma­ codylomorpha (71%). When birds were pruned trix, the unpruned 50% majority-rule summary away, maniraptorans were still in a weak poly­ tree of 500 bootstrap replicate trees had minimal tomy with Longisquama and were only ambigu­ structure, so we pruned away taxa in various ously associated with nonmaniraptoran theropods ways to ask specific questions with smaller sets and Crocodylomorpha (not shown). Even when of taxa that retained representatives of various Longisquama was pruned away from Figure 9, the combinations of major groups. clade of birds and core maniraptorans was weakly First, to explore the relationships among basal supported (66%) and was still in a polytomy with birds and representative taxa of the most bird­ a clade of nonmaniraptoran theropods (65%) and like core maniraptoran groups (oviraptorosaurs, Crocodylomorpha (72%) (Fig. 10). Although Aves

OM66_Text.indd 18 3/31/09 5:54:56 PM origin of birds 19

Table 2. Statistics from reanalysis of the 46-taxon × 208-character matrix of Clark et al. (2002), with the lengths of most- parsimonious trees (MPTs) when certain theropod taxa were constrained to be within Aves as defined here.

Kishino-Hasegawa testa

Length Range of P (one- Missing Number of Difference standard tailed Constraint n data (%) of MPTs MPTs in length deviations t tests)

None — 52 356 626 Shuvuuia 1 347 626 0 — — — Alvarezsauridae 4 61 644 626 0 — — — Microraptor 1 50 10,000 627 1 8.0–8.2 0.125 0.451 Troodontidae 6 55 10,000 627 1 8.0 0.126 0.450 Dromaeosauridae 11 59 10,000 627 1 8.0 0.126 0.450 Dromaeosauridae and 17 58 10,000 627 1 8.0 0.126 0.450 Troodontidae Caudipteryx 1 61 2,424 629 3 8.0–8.2 0.348 0.368 Therizinosauroidea 3 66 10,000 629 3 7.7–8.2 0.373 0.363 Oviraptorosauria 9 56 2,416 629 3 8.0–8.2 0.368 0.368 Oviraptorosauria, 26 57 1,576 631 5 6.7 0.745 0.229 Troodontidae, Dromaeosauridae Maniraptora (no Aves) 33 58 2,048 635 9 7.5 1.193 0.117 Ornithomimosauria 5 52 10,000 652 26 8.9–9.1 2.901 (0.0021) Maniraptoriformes 38 58 10,000 658 32 10.6 3.029 (0.0014) (no Aves)

Notes: n is the number of taxa that were constrained. The consistency index ranged from 0.39 to 0.41; the retention index ranged from 0.66 to 0.69. Comparisons are ordered by the differences in the number of steps in tree length from no constraint (626). Of the 13 comparisons, the first 11 were unable to reject the hypothesis that the trees are sample estimates of the same phylogeny. a Results in Tables 2, 3, and 4 are reported as the range of standard deviations and the average t and P values for five tests between randomly selected topologies between MPTs for no constraint and randomly selected topologies with the stated backbone con­ straint. Kishino-Hasegawa tests are nonparametric likelihood ratio tests. They give two-tailed tests of differences in topology. We report one-tailed P values, which indicate nonrejection of the hypothesis of no difference in the first 11 comparisons. The results also indicate inability to make proper allowance for a-posteriori selection of the most-parsimonious trees in the last two comparisons, for which the two-tailed P values were less than twice the value required to indicate rejection of the hypothesis of no difference (0.025). Interpretation is also complicated by bias in the test toward finding significant valuesG ( oldman et al. 2000). We therefore put values <0.025 in parentheses.

was weakly supported (53%) on this tree, interre­ (Fig. 12). This clade of birds and Longisquama was lationships within the clade of birds and manirap­ in a polytomy with Crocodylomorpha (70%), basal torans were still largely unresolved. When 35 taxa saurischians, and nonmaniraptoran theropods. remained, representing all major groups proposed When Longisquama was also pruned away, and 23 as possible sister-groups of birds (oviraptorosaurs, taxa remained, Aves was strongly supported (93%) troodontids, dromaeosaurs, alvarezsaurids, non­ but was still in a polytomy with Crocodylomorpha maniraptoran theropods, basal archosaurs, and (71%), basal saurischians, and nonmaniraptoran crocodylomorphs), the had little struc­ theropods (Fig. 13). ture (Fig. 11). A large polytomy of archosaurs more When various alvarezsaurids, dromaeosaurs, derived than the basal forms Proterosuchus, Eupark- troodontids, or oviraptorosaurs were constrained eria, and Erythrosuchus is strongly supported (98%), to be within Aves as defined here (as inF igs. 1B, 3F), but neither Aves nor Theropoda is recovered, and MPTs were 3–7 steps longer than unconstrained birds are not unambiguously closer to theropods MPTs, with a jump to 17 steps for therizinosau­ than to other archosaurs. When we asked about roids and 23 for ornithomimosaurs (Table 3). remaining relationships when the maniraptorans Except for the Maniraptora and Maniraptori­ had been pruned away, however, birds were as­ formes as a whole, Kishino-Hasegawa tests in­ sociated with Longisquama in a clade (55%) in dicated that the unconstrained and constrained which interrelationships were largely unresolved trees were not statistically different.

OM66_Text.indd 19 3/31/09 5:54:57 PM 20 ORNITHOLOGICAL MONOGRAPHS NO. 66

Fig. 9. A 50% majority-rule summary of 500 bootstrap replicates of most-parsimonious trees calculated from a matrix of 79 taxa including theropods, birds, nontheropod dinosaurs, and other archosaurs, pruned to 20 taxa. Pruning reveals a weak clade of birds, core maniraptorans, and Longisquama (58%), all in a polytomy with non­ maniraptoran theropods (66%) and crocodylomorphs (71%).

Fig. 10. As in Figure 9, but with Longisquama also pruned away. The clade of birds and core maniraptorans has gained support (65%), but it is still not unambiguously closer to other theropods (66%) than to the crocodylo­ morphs (72%). Even with the removal of Longisquama, the relationship of the clade of birds and core manirap­ torans to other theropods is ambiguous.

OM66_Text.indd 20 3/31/09 5:54:58 PM origin of birds 21

Fig. 11. As in Figure 9, pruned to 35 taxa from 79, representing all major proposed sister groups of birds. There is a large polytomy of archosaurs more derived than the basal forms Proterosuchus, Euparkeria, and Erythrosuchus (98%), in which neither Aves nor Theropoda are recovered and birds are not unambiguously closer to theropods than to other archosaurs.

Next, we repeated our analysis with backbone of the MPTs, in that . The constrained and constraints for the five topologies representing the unconstrained MPTs did not differ significantly. five hypotheses for the origin of birds (Table 4). The relative lengths of the MPTs, given their Kishino-Hasegawa tests indicated that the early- backbone constraints, supported this conclusion archosaur, crocodylomorph, and BMT hypoth­ (Table 4). The differences were 14, 18, and 27 eses were most compatible with the topology steps, respectively.

OM66_Text.indd 21 3/31/09 5:54:59 PM 22 ORNITHOLOGICAL MONOGRAPHS NO. 66

Fig. 12. As in Figure 11, but with the Maniraptora pruned away. Birds are associated with Longisquama in a weak clade (55%). That clade is not unambiguously associated with theropods.

Fig. 13. As in Figure 12, but with both maniraptorans and Longisquama pruned away. The sister group of birds is still unresolved. Birds are not clearly more closely related to theropods than to crocodylomorphs.

OM66_Text.indd 22 3/31/09 5:55:01 PM origin of birds 23

Table 3. Statistics from analysis of our 79-taxon × 221-character matrix with lengths of most-parsimonious trees (MPTs) when certain theropod taxa were constrained to be within Aves as defined here.

Kishino-Hasegawa testa

Length Range of P (one- Missing Number of Difference standard tailed Constraint n data (%) of MPTs MPTs in length deviations t tests)

None — 39 360 1,355 Shuvuuia 1 23 2,169 1,358 3 11.5–13.2 0.245 0.404 Alvarezsauridae 2 43 4,114 1,358 3 12.3–12.6 0.239 0.405 Microraptor 1 41 2,106 1,360 5 11.6–12.5 0.417 0.339 Dromaeosauridae 7 42 2,085 1,360 5 11.8–12.4 0.410 0.339 Caudipteryx 1 37 444 1,360 5 10.1–11.1 0.463 0.267 Troodontidae 5 50 305 1,362 7 11.2–13.2 0.581 0.281 Dromaeosauridae and 12 46 42 1,362 7 10.0–11.3 0.669 0.252 Troodontidae Oviraptorosauria 9 50 192 1,362 7 10.5–11.5 0.643 0.261 Oviraptorosauria, 21 47 315 1,371 16 10.9–12.2 1.393 0.083 Dromaeosauridae, Troodontidae Therizinosauroidea 4 59 26 1,372 17 13.4–14.1 1.239 0.109 Ornithomimosauria 4 40 2 1,378 23 12.9–13.5 1.738 0.042 Maniraptora (no Aves) 27 49 7,230 1,387 32 13.1–15.0 2.249 (0.013) Maniraptoriformes 31 48 43 1,392 37 15.1–15.8 2.402 (0.009) (no Aves)

Notes: n is the number of taxa that were constrained. The taxa in the full matrix are examples from all archosaur groups proposed to be ancestors of birds, including maniraptoran and nonmaniraptoran theropod dinosaurs, crocodylomorphs, and other nonthero­ pod archosaurs. The consistency index ranged from 0.23 to 0.24. The retention index ranged from 0.62 to 0.64. Comparisons are ordered by differences in the number of steps in tree length from no constraint (1,355). a See note a in Table 2 for interpretation of results.

Alternative Analysis of the New Matrix archosaur Proterosuchus for rooting. As in the primary analysis, unpruned 50% majority-rule When we analyzed the alternative matrix, with consensus trees displayed minimal structure. Af­ its 79 taxa and 242 characters, without topologi­ ter bootstrapping and pruning, the topology of cal constraints (using PAUP*), we recovered 104 50% majority-rule consensus trees did not differ MPTs of 1,360 steps. Again, we used the basal appreciably from those reported in Figures 9–13.

Table 4. Statistics from analysis of our 79 taxon × 221 character matrix with lengths of most-parsimonious trees (MPTs) when taxa were constrained to match hypotheses for the origin of birds (BMT = birds are maniraptoran theropod dinosaurs).

Kishino-Hasegawa testa

Range of Number Length Difference standard P (one-tailed Constraint of MPTs of MPTs in length deviations t tests)

None 360 1,355 BMT hypothesis 4,624 1,382 27 14.2–15.1 1.864 (0.032) Neoflightless-theropod 52 1,391 36 16.5–17.5 2.137 (0.017) hypothesis Early-archosaur hypothesis 12 1,369 14 10.8–11.7 1.237 0.109 Crocodylomorph hypothesis 379 1,373 18 10.9–12.1 1.552 0.061 Birds evolved twice 56 1,441 86 16.1–17.3 5.202 (0.00005)

a See note a in Table 2 for interpretation of results.

OM66_Text.indd 23 3/31/09 5:55:02 PM 24 ORNITHOLOGICAL MONOGRAPHS NO. 66

Fig. 14. A 50% majority-rule summary of 500 bootstrap replicates of most-parsimonious trees calculated from an alternative analysis to the analyses reported in Figures 9–13, using 242 instead of 221 characters, and pruned as in Figure 11. Twenty-one characters for questionable homologies, previously turned off, were turned on for this alternative analysis. They were scored as unknown (“?”) where appropriate. Note that the topologies of Figures 11 and this tree are the same and differ only slightly in bootstrap support.

For example, Figure 11, based on the 79 × 221 bootstrap values, 11 changes were shifts of only matrix, and Figure 14, based on the alternative one or two percentage points, and the largest 79 × 242 matrix, both have the lowest level of change was a shift of six percentage points for pruning and can be used for the most general the Ornithomimosauria. Similar comparisons comparisons among alternative hypotheses. The were made for Figures 9, 10, 12, and 13. The only topologies do not differ, and only minor differ­ topological differences were in the comparison ences in bootstrap support were apparent. Of 13 with Figure 10, where the weakly supported Aves

OM66_Text.indd 24 3/31/09 5:55:03 PM origin of birds 25

was not recovered in the alternative analysis, cladistic methods, insufficient tests of primary ho­ and in the comparison with Figure 12, where the mology statements, lack of statistical evaluation, polytomy of basal birds and Longisquama was use of verificationist arguments, and introduction weakly resolved in favor of a sister-group rela­ of ad-hoc auxiliary hypotheses. We conclude that tionship between Aves and Longisquama. Again, the BMT hypothesis has not been critically tested differences in bootstrap values were negligible. and that the adoption of a verificationist approach and the introduction of ad-hoc auxiliary hypoth­ Discussion eses have impaired its testability. Unjustifiable assumptions of homology incorporated This monograph is a response to defenders of into data matrices.—The most glaring example of the BMT hypothesis who contend that the BMT this problem is the coding of avian and theropod hypothesis is correct because no more parsimoni­ manual, carpal, and tarsal characters as if they were ous cladistic analysis has been offered by its crit­ homologous, despite the ambiguity of the data, ics (e.g., Padian 2001b, Prum 2003). We agree that and despite the assumption this coding entails that questions about the origin of birds are best viewed the BMT hypothesis is correct a priori (Martin et al. in a comparative phylogenetic framework, but 1980; Martin 1983, 1991; Martin and Stewart 1985; logically a hypothesis can be assessed even in Burke and Feduccia 1997; Feduccia 1999; Wagner the absence of a clearly preferred alternative hy­ and Gauthier 1999; Feduccia and Nowicki 2002; pothesis. In any case, here we have taken both ap­ Kundrát et al. 2002; Larsson and Wagner 2002; Ga­ proaches. Our study had two goals: (1) to assess lis et al. 2003, 2005; Mayr et al. 2005, 2007; Vargas whether the BMT hypothesis is overwhelmingly and Fallon 2005a, b; Welten et al. 2005; Appendix supported by published cladistic analyses and (2) 3). This practice conflicts with the recommenda­ to evaluate the BMT hypothesis in the context of tions of Rieppel and Kearney (2002) and Kearney alternatives within a comparative phylogenetic and Rieppel (2006) about ways to avoid the dan­ framework, including the proposition that some gers of incorporating unjustifiable assumptions of maniraptorans belong within Aves. Our inten­ homology into data matrices. Unfortunately, these tion in pursuing the second goal was to allow for problems persist even in the improved matrices possible refutation of the BMT hypothesis. Our of the Theropod Working Group, like the CNM review and reanalysis of a representative BMT matrix of Clark et al. (2002). For example, Clark matrix do not support the contention of BMT et al. (2002) used manual characters scored for advocates either that the hypothesis has been both birds and theropods, thereby assuming their tested with cladistics or that it is overwhelmingly homology. In addition to the importation of bias supported. Our cladistic and statistical analyses through unjustified assumptions of homology, of our new data set indicate that several predic­ the current BMT literature, like much literature in tions derived from the BMT hypothesis are not morphological cladistics, does not provide rigor­ supported and that alternatives to the BMT are at ous analysis of characters. Ambiguities in the data least equally viable. Altogether, three hypotheses that complicate assessment of primary homolo­ for the origin of birds—the BMT, early-archosaur, gies are usually not addressed in the current BMT and crocodylomorph hypotheses—are most com­ literature (see Appendix 3 for detailed analyses of patible with currently available evidence. some problematic characters). The congruence of other characters is some­ Has the BMT Hypothesis Been Tested? times offered as justification for such assump­ tions: birds are theropods because they share If the BMT hypothesis were as well supported other characters with theropods and, therefore, as its proponents claim, a review of the support­ must have the same digital identities as theropods ing literature and a reanalysis of a representative (e.g., Wagner and Gauthier 1999, Makovicky and matrix from that literature should reveal no seri­ Dyke 2001, Padian 2001b). This reasoning is cir­ ous problems, but our review of the literature and cular. Synapomorphies are invoked to defend the reanalysis of data from Clark et al. (2002) revealed hypothesis; the hypothesis is invoked to defend (at least) seven problems, which we discuss be­ the synapomorphies. low: unjustifiable assumptions of homology in­ Inadequate taxon sampling.—Like the CNM ma­ corporated into data matrices, inadequate taxon trix, all analyses we could find that were cited as sampling, insufficiently rigorous application of supporting the BMT hypothesis failed to include

OM66_Text.indd 25 3/31/09 5:55:04 PM 26 ORNITHOLOGICAL MONOGRAPHS NO. 66

nondinosaurian taxa. Such analyses cannot test our results on 50% majority-rule trees. The combi­ the BMT hypothesis, because they do not allow nation of bootstrapping and pruning used in our for the possibility of its refutation. At best, they reanalysis of the CNM matrix, and in the analysis are attempts to examine the interrelationships of of our new matrix, revealed structure in the data birds and theropods, once Aves is assumed to be and in subsets of the data for which some esti­ nested inside Theropoda. mates of statistical support could be made. Insufficiently rigorous application of cladistic As was the case in our reanalysis of the CNM methods.—Clark et al. (2002) used the program matrix, high resolution on a strict-consensus tree NONA (Goloboff 1999) to perform a heuristic can prove to be spurious when bootstrapping and search for unrooted MPTs. They used branch pruning are used to evaluate support for the con­ swapping extended to suboptimal trees up to sensus topology. Our reanalysis revealed that the 10% longer, retrieved MPTs, and then calculated structure of the strict consensus tree of Clark et al. consistency and retention indices. They then com­ (2002) has minimal statistical support. A transfer puted a strict consensus of the MPTs and rooted of bootstrap values from the 50% majority-rule it, in this case with the noncoelurosaurian thero­ summary of the bootstrap replicates (Fig. 7) to pods Allosaurus and Sinraptor (Clark et al. 2002, the strict consensus tree reported by Clark et al. fig. 2.2A). No further tree evaluation, such as the (2002) would have revealed that almost none of bootstrap procedure, was reported, and the strict- the clades reported on that tree have ≥50% boot­ consensus tree was taken as a well-supported re­ strap support. Interestingly, although Clark et al.’s construction of the phylogeny of the ingroup taxa. (2002) strict consensus tree recovers Aves deeply Except in a few recent papers (e.g., Marya´nska et nested within Maniraptora, bootstrapping com­ al. 2002), the methods used by Clark et al. (2002) bined with pruning reveals that the relationships are typical of the methods employed throughout among the maniraptoran and avian taxa are actu­ the BMT literature. ally unresolved (Figs. 7 and 8), leaving open the As emphasized by Felsenstein (2004) and oth­ possibility that some maniraptoran taxa belong ers, proper rigor requires searching for relation­ within Aves. Given the importance of the rela­ ships supported by a known proportion of trees tionship between birds and maniraptorans in the and determination of what subset of relation­ debate about the origin of birds, this previously ships among the population of trees is well sup­ overlooked ambiguity is significant. This lack of ported, rather than searching for the optimal tree resolution indicates uncertainties in the cladistic and then assuming that it is the correct phylog­ support for the BMT hypothesis that have rarely eny from which character evolution can reliably been acknowledged. be reconstructed. Failure to assess the relation­ Of course, our methods are not without their ships between the evidence (the matrix) and the own deficiencies. Although bootstrapping was hypothesis (the cladogram) with procedures like used in our study, bootstrap values only give an the bootstrap, and the assumption that a strict indication of the strength of support for a branch consensus tree accurately reflects supported re­ and should, therefore, be interpreted cautiously lationships, are serious deficiencies in much of (Swofford et al. 1996). Although Swofford et al. the literature in (Swofford et al. 1996, (1996) recommended weighting characters and Felsenstein 2004, Jenner 2004). creating step matrices as ways to incorporate Pruning used in combination with bootstrap­ independent data, we did not use these options. ping can further enhance comparisons between We felt that selection of weights and step matri­ the evidence and the hypotheses being evaluated. ces was as likely to add bias to the analysis as to Through successive pruning, one can attempt reduce it. Neither did we evaluate the relation­ to determine how specific taxa affect support ship between character conflicts and missing data for branches as revealed by bootstrap analyses (Kearney 2002, Kearney and Clark 2003). (Wilkinson 1996). Therefore, pruning is useful in Insufficient tests of primary homology statements.— exploratory analyses of subsets of the data and al­ Given its problems in dealing with homoplasy lows for more sensitive estimation of support for (e.g., Gosliner and Ghiselin 1984; Carroll 1988; branches in 50% majority-rule trees (Wilkinson Carroll and Dong 1991; Livezey 1998, 2003; Fe­ 1996, 2003; Steel and Penny 2000; Cranston and duccia 1999; Wiens et al. 2003), cladistics alone Rannala 2007; Holder et al. 2008). We used boot­ should not be relied upon to test phylogenetic strapping and successive pruning and reported hypotheses, and “tests of congruence” should

OM66_Text.indd 26 3/31/09 5:55:04 PM origin of birds 27

not be relied upon to evaluate primary homology provided that the observation is repeatable—and, statements (Rieppel and Kearney 2002, Kearney thus, not attributable to a flawed experiment or and Rieppel 2006). Independent testability of pri­ faulty observation (see Popper [2002] on the ne­ mary homology statements should be grounded cessity of repeatability)—logic dictates that even in comparative anatomical research (Rieppel and a single observation could be enough to falsify Kearney 2002, Kearney and Rieppel 2006). Even the BMT hypothesis if it contravenes a prediction though a cladistic analysis does not actually test logically deducible from that hypothesis. Con­ a hypothesis—it simply produces the most par­ sequently, the verificationist approach is neither simonious phylogeny, given the assumptions in­ logically nor empirically sound. It has discour­ corporated into the analysis a priori—many have aged critical evaluation of the BMT hypothesis claimed that the BMT hypothesis has been tested and encouraged the uncritical practice of looking with cladistics (e.g., Padian and Chiappe 1998; only for supporting evidence. Padian 2001b; Prum 2002, 2003). Introduction of ad-hoc auxiliary hypotheses.—An Lack of statistical evaluation.—Recent efforts to ad-hoc auxiliary hypothesis is one that has been introduce statistical methods into phylogenetic formulated for the specific purpose of restoring analyses (e.g., Felsenstein 2001, 2004) offer ad­ agreement between a hypothesis and falsifying ditional techniques for strengthening evaluation observations; it serves no independent explana­ of phylogenetic hypotheses. Unfortunately, the tory function and does not entail any significant, BMT literature continues to interpret cladistic re­ independently testable implications (e.g., Hempel sults without statistical evaluation. 1966, Popper 2002). Although ad-hoc auxiliary Verificationist arguments.—The BMT hypoth­ hypotheses are often used to protect favored hy­ esis has been claimed to be overwhelmingly sup­ potheses (Kuhn 1970), and although they may ported by the many synapomorphies shared by be empirically valid, they actually interfere with birds and theropods (e.g., Padian and Chiappe testability (Hempel 1966, Popper 2002) by in­ 1998; Chiappe and Dyke 2002; Paul 2002; Prum creasing the range of observations with which 2002, 2003; Xu et al. 2003; Xu 2006). The hypoth­ a hypothesis is compatible. If the introduction esis is inferred to be correct because of the vast of ad-hoc auxiliary hypotheses were considered quantity of data held to support it. This argument legitimate, they could be used to explain away is logically problematic. The number of confirma­ all falsifying observations, rendering a favored tory observations that a hypothesis can marshal is hypothesis immune to any criticism. Repeatedly not necessarily relevant to whether the hypothesis obtained observations that contradict a hypoth­ is correct, and the history of science offers numer­ esis should be accepted as falsifying observations ous examples of “overwhelmingly supported” rather than explained away. theories that were later refuted (Hempel 1966; The “frame-shift hypothesis” of Wagner and Popper 1998, 2002). The risk is that only support­ Gauthier (1999) is an example of an ad-hoc auxil­ ing evidence will be recognized, while contradic­ iary hypothesis. The tridactyl manus of neoteta­ tory evidence is ignored or explained away. Kluge nurine theropods appears to be composed of the (1997, 1999, 2001) has repeatedly urged the adop­ first, second, and third digits of the primitively tion of a falsificationist, rather than a verification­ pentadactyl archosaurian manus (e.g., Gauthier ist, research program for cladistics. 1986, Wagner and Gauthier 1999; see Appendix 3), Logic aside, criticism of putative synapomor­ which implies that, if Aves is nested within phies shared by birds and theropods has per­ Theropoda, the digits of the tridactyl avian manus sisted. In particular, inferences of homologies for should be digits I, II, and III. Embryological data characters of the carpus, manus, and tarsus have indicate, however, that the digits of the avian been questioned (e.g., Burke and Feduccia 1997; manus are in fact the second, third, and fourth Feduccia 1999, 2002; Feduccia and Nowicki 2002; digits of the primitively pentadactyl archosau­ Martin 2004; Feduccia et al. 2005, 2007; see Ap­ rian manus (e.g., Feduccia 2002, Feduccia and pendix 3). One prominent reply to these concerns Nowicki 2002, Kundrát et al. 2002, Larsson and is to appeal to the reasoning outlined above (e.g., Wagner 2002; see Appendix 3). This repeatedly Wagner and Gauthier 1999, Makovicky and Dyke obtained observation contradicts a key prediction 2001, Padian 2001b), arguing that the preponder­ of the BMT hypothesis and is logically sufficient ance of confirming data allows any given falsi­ to refute it. To explain away this falsifying obser­ fying observations to be discounted. However, vation, Wagner and Gauthier (1999) proposed

OM66_Text.indd 27 3/31/09 5:55:04 PM 28 ORNITHOLOGICAL MONOGRAPHS NO. 66

that a homeotic frame-shift affecting digital iden­ aberrant distribution of feathers in this clade; and tity must have occurred during the evolution of (4) regional climatic variation produced varia­ neotetanurine theropods (see Appendix 3). tion in distribution of feathers among compsog­ Although some data suggest that homeotic nathids. frame-shifts such as those postulated by Wagner Like the frame-shift hypothesis of Wagner and Gauthier (1999) can occur in evo­ and Gauthier (1999), these four hypotheses were lution, and perhaps occurred during theropod introduced for the explicit purpose of restoring evolution (e.g., Vargas and Fallon 2005a, b, and agreement between a prediction that is a key references therein), these data do not alter the component of the BMT hypothesis and an unam­ logical status of the frame-shift hypothesis as an biguously falsifying observation. Although any ad-hoc auxiliary hypothesis (for a review of the one of these four hypotheses could be correct, data, see Appendix 3). It was introduced for the none of the four explains previously unexplained explicit purpose of restoring agreement between data in theropod or avian evolution or generates predictions of the BMT hypothesis and repeat­ any new, significant, testable predictions that edly obtained falsifying observations. It explains might enhance the overall testability of the BMT no previously unexplained data in theropod or hypothesis. Like the frame-shift hypothesis, they avian evolution and generates no new, signifi­ are ad-hoc auxiliary hypotheses that restrict the cant, testable predictions that might enhance the range of observations that could contravene the overall testability of the BMT hypothesis. As an BMT hypothesis and interfere with its testing. ad-hoc auxiliary hypothesis, the frame-shift hy­ pothesis restricts the range of observations that Evaluation of Alternative Hypotheses could contravene the BMT hypothesis and inter­ feres with its testing. Although we think that the BMT hypothesis Göhlich and Chiappe (2006) provide another has not been tested and is not as overwhelm­ example. Feathers are hypothesized to be a sy­ ingly supported as has been claimed, it is not, napomorphy of at least the Coelurosauria (e.g., for those reasons, necessarily incorrect. Analysis Chiappe and Dyke 2002, Norell and Xu 2005), be­ of our new matrix, however, which allows for cause of the identification of fibrous structures in evaluation within a comparative framework of basal coelurosaurs like the compsognathid Sino- the BMT hypothesis and four alternative hypoth­ sauropteryx (Currie and Chen 2001) as feathers. eses for the origin of birds (Fig. 3), and review of The prediction that any fossil of a coelurosaurian the literature, indicate (1) that several predictions taxon with large tracts of well-preserved integu­ derivable from the BMT hypothesis are not sup­ ment should display feathers can be logically ported; (2) that some maniraptorans may belong deduced from this hypothesis. Consequently, the within Aves, which potentially supports the three discovery of a coelurosaurian fossil with well- alternatives to the BMT hypothesis that incor­ preserved integument lacking feathers would be porate this topology (the neoflightless-theropod logically sufficient to falsify the hypothesis that hypothesis, the early-archosaur hypothesis, and all coelurosaurs were feathered, a key component the crocodylomorph hypothesis); (3) that avian of the BMT hypothesis itself. The compsognathid status for even some maniraptorans weakens Juravenator, described by Göhlich and Chiappe support for both the BMT hypothesis and the (2006), is exquisitely preserved, and the type neoflightless-theropod hypothesis; and (4) that, material includes large tracts of well-preserved of the alternatives to the BMT hypothesis, the integument. Nevertheless, no feathers are pre­ early-archosaur and crocodylomorph hypotheses served in the Juravenator type material. Göhlich are equally compatible with currently available and Chiappe (2006) explained away this falsify­ evidence. We expand on these points below. ing observation by postulating four distinct ad-hoc Are predictions of the BMT hypothesis supported by auxiliary hypotheses: (1) feathers evolved repeat­ our results?—We derived several predictions from edly in Coelurosauria, accounting for discrepan­ the BMT hypothesis and checked to determine cies in distribution among well-preserved whether they were supported by our analyses (see coelurosaurian taxa; (2) some coelurosaurs, like Table 5). For example, the BMT hypothesis holds Juravenator, secondarily lost their feathers; (3) that Aves is deeply nested within the Maniraptora unspecified and unknowable autapomorphic and that maniraptoran clades like the Oviraptoro­ features in compsognathid produced , Dromaeosauridae, and Troodontidae are

OM66_Text.indd 28 3/31/09 5:55:05 PM origin of birds 29

Table 5. five predictions of the BMT (birds are maniraptoran theropod dinosaurs) hypothesis, along with indications from the present analysis of the Clark et al. (2002) matrix for coelurosaurs and of our matrix, in which all hypotheses can be evaluated simultaneously.

Indications of support Indications of from the Clark et al. support from analysis Prediction of the BMT hypothesis (2002) reanalysis of our matrix

The lengths of most-parsimonious trees and those of trees No (Table 2) No (Table 3) in which various nonavian maniraptorans are constrained to be within Aves as defined here will differ statistically. In , Aves will be nested within the Maniraptora. Ambiguous (Fig. 8) Ambiguous (Figs. 9–11) A clade of birds and maniraptorans will be nested within Yes (Fig. 8) No (Figs. 9–11) the Theropoda. When potential sister groups for birds among early — No (Fig. 10) archosaurs are pruned away, a clade of birds and maniraptorans will be clearly associated with theropods. When maniraptorans are pruned away, birds will be clearly — No (Figs. 12 and 13) associated with theropods instead of with early archosaurs or crocodylomorphs.

nonavian. Consequently, in both the reanalysis of but interrelationships were ambiguous (Fig. 8). the CNM matrix and the analysis of our new ma­ Similarly, our analysis of our matrix showed trix, the lengths of unconstrained MPTs and trees that birds and maniraptorans are associated; but in which various maniraptorans are constrained again, interrelationships are ambiguous (Figs. to be inside Aves should differ statistically. When 9–11). Although they do not establish it, both the backbone-constraint options in PAUP* and results are compatible with the possibility that the Kishino-Hasegawa tests were used, and the some maniraptorans belong within Aves. warnings of Goldman et al. (2000) about such ap­ A third prediction of the BMT hypothesis is plications were accommodated, this prediction that a clade of birds and maniraptorans should failed both with the CNM data (Table 2) and with be nested within the Theropoda, a relationship ours (Table 3). The concerns of Goldman et al. supported by our reanalysis of the CNM matrix (2000) were about bias toward finding statistical (Fig. 8), where the only taxa used were birds and differences, not toward the small, nonsignificant theropods, but not by the analysis of our matrix, differences that we found. These results conflict where the relationships among birds and other with what would be expected if the BMT hypoth­ archosaurs are ambiguous (Figs. 9–11). esis were correct. They indicate that a topology A fourth prediction is that, when potential sis­ in which at least some maniraptorans are actu­ ter groups for birds among early archosaurs are ally birds cannot be rejected as statistically sig­ pruned away (and the position of Longisquama nificantly worse than the standard BMT topology. recovered in other trees is thereby discounted; Clearly, the possibility that some maniraptoran see Figs. 9 and 12), a clade of birds and manirap­ groups are birds needs further testing and should torans will be unambiguously associated with not be dismissed. See Corfe and Butler (2006) for theropods. We found that this clade was only am­ a similar approach to determining whether alter­ biguously associated with theropods in a poly­ native hypotheses are statistically different from tomy with other archosaurs (Figs. 10 and 13). the preferred hypothesis. A fifth prediction is that, when maniraptorans A second prediction of the BMT hypothesis is are pruned away, birds should still be clearly as­ that Aves should be nested within the Manirap­ sociated with more basal theropods, which have tora, as shown in the strict consensus tree of Clark been repeatedly identified as having numerous et al. (2002, fig. 2.2). This position is consistent birdlike characters (e.g., Padian 2001b, Paul 2002). with the generally accepted phylogeny for the Again, however, birds were only ambiguously as­ Archosauria as a whole (Fig. 2). Our reanalysis sociated with theropods in a polytomy of higher of the CNM matrix retrieved Archaeopteryx and archosaurs (Figs. 12 and 13). In summary, when Confuciusornis as associated with maniraptorans, statistical tests, bootstrap support, and pruning

OM66_Text.indd 29 3/31/09 5:55:05 PM 30 ORNITHOLOGICAL MONOGRAPHS NO. 66

are applied to both the standard CNM matrix of Sinornithosaurus, are expected to be morphologi­ Clark et al. (2002) and our more inclusive matrix, cally more similar to flying birds than later, more major ambiguities are revealed about the strength derived forms, like Deinonychus and Velociraptor of support for the BMT hypothesis, and BMT pre­ (Paul 2002). This prediction is corroborated by dictions are not supported. Consequently, we the morphology of these taxa (Paul 2002). conclude that the BMT hypothesis is not only un­ The cladistic analysis of Mayr et al. (2005) sup­ tested but also not more clearly supported than ports the placement of dromaeosaurs within Aves; two of the alternatives when evaluated within a in their cladogram, dromaeosaurs (and troodon­ comparative phylogenetic framework allowing tids) are the sister group of Archaeopteryx and the for its possible refutation. basal bird Rahonavis, and Confuciusornis is nested Are some maniraptorans actually birds more de- within a clade of troodontids and dromaeosaurs rived than Archaeopteryx?—Our analyses of both (Mayr et al. 2005, fig. 4). Contrary to Mayr et al. the CNM matrix and our new matrix support the (2005), the simpler interpretation of this topology possibility that some maniraptorans are birds is that it supports avian status for dromaeosaurs more derived than Archaeopteryx, in the sense of (and troodontids), rather than supporting the au­ Figures 1B and 3F. This is a major component of thors’ interpretation of avian diphyly. three of the alternatives to the BMT hypothesis, The situation of oviraptorosaurs is less clear. including the two supported by our results, the Derived oviraptorosaurs, like Citipati and Inge- early-archosaur and crocodylomorph hypothe­ nia, possess a number of advanced avian apo­ ses. Our results suggesting avian status for some morphies: fused prefontals, reduced maxillae, maniraptorans indicate that these alternatives extensively pneumatized narial region, the shape should not be rejected at this time. The potential of the lacrimal (“reverse C-shaped,” as in, e.g., avian status of some maniraptorans should be ex­ Confuciusornis; see Martin et al. 1998b), contral­ plored with independent evidence. ateral communication between at least some With respect to Dromaeosauridae, a strong case tympanic diverticulae, fusion of the articular and has been made for the avian status of Microraptor, surangular, articular surface for quadrate with which has been described as a basal dromaeosaur development of either a lateral or a medial pro­ (Xu et al. 2000, 2003; Hwang et al. 2002). Microrap- cess or both, pneumatic presacral vertebrae, more tor has multiple characters more derived toward than five sacrals, ossified uncincate processes, modern birds than those of Archaeopteryx, particu­ ossified sternal plates, costal facets on sternum, larly in the pectoral girdle: posterior cervicals with sternum with lateral process, and anterior mar­ carotid processes, more than five sacrals, ossified gin of sternum grooved anterolaterally for re­ uncinate processes, ossified sternal plate (rather ception of coracoids, among others (Elzanowski than separate plates), costal facets on sternum, 1999, Marya´nska et al. 2002, Paul 2002, Lü et al. sternum with lateral process, anterior margin of 2005). The morphology of basal forms like Cau- sternum grooved anterolaterally for reception of dipteryx is more ambiguous and more primitive, coracoids, /coracoid angle acute, antitro­ but the three characters cited by Ji et al. (1998) chanter present on , and tibiotarsus present, and Witmer (2002) in arguing that Caudipteryx is among others (Xu et al. 2000, 2003; Czerkas et al. not a bird are either incorrect or ambiguous: (1) 2002; Hwang et al. 2002; Feduccia et al. 2005, 2007; the quadratojugal is not sutured to the quadrate Burnham 2007). Its fore- and hind-limb airfoils (Ruben and Jones 2000); (2) whether the quadra­ are indistinguishable from those of Archaeopteryx tojugal and squamosal made contact is unclear; (Longrich 2006). Burnham (2007) argues that the and (3) the presence of an obturator process is basal dromaeosaur Bambiraptor is also a bird, and not decisive because an obturator process is also Paul (2002) summarizes the character evidence present in some Mesozoic birds, such as Con- supporting the avian status of all dromaeosaurs. cornis (Sanz et al. 2002). Some of the characters If basal dromaeosaurs belong within Aves, then, of Caudipteryx cited as purported plesiomorphies given avian and dromaeosaur mono­ by Witmer (2002) and Chiappe and Dyke (2002), phyly, all dromaeosaurs must belong within Aves such as the nonavian position of the scapula and (Czerkas et al. 2002; Feduccia 2002; Paul 2002; the nonavian morphology of the coracoid, would Feduccia et al. 2005, 2007). If dromaeosaurs are be expected if Caudipteryx were a . birds, most of which had lost the ability to fly, The significance of some of the other characters more-basal dromaeosaurs, like Microraptor and they cite is not clear (e.g., Caudipteryx has a deep

OM66_Text.indd 30 3/31/09 5:55:05 PM origin of birds 31

mandibular fossa, unlike most Mesozoic birds, to difficulties. Dramatic restructuring of the skel­ but so does Confuciusornis). Caudipteryx is mor­ eton would be required to derive even as bird­ phologically no more primitive than Archaeop- like an as Bambiraptor from something like teryx in its tooth morphology, hallucal position, Archaeopteryx. The tail would have to be elon­ and quadrate morphology; it is more similar gated, the elongate prezygapophyses reacquired, morphologically than is Archaeopteryx to ornithu­ various braincase and otic-region characters lost, rines in its possession of a ventral foramen mag­ and the tooth morphology and the implantation num and ossified sternal plates (Feduccia 1999, system changed. To derive something like Dei- Paul 2002, Feduccia et al. 2005). Some evidence nonychus or a troodontid would require further indicates that other oviraptorosaurs (like Citipati) changes, but flight loss is known to cause dra­ were birds, and a few cladistic analyses have re­ matic restructuring of the skeleton in birds, often covered all oviraptorosaurs within Aves (e.g., Lü obscuring phylogenetic relationships (Livezey et al. 2002, Marya´nska et al. 2002). 1998, 2003; Feduccia 1999, 2002; Paul 2002). The If the preceding arguments are granted, a re­ Dromornithidae (Neornithes: ) of­ gress is opened: how many other maniraptoran fer a particularly striking example of how mor­ taxa might belong within Aves? A case for the phological restructuring associated with flight avian status of Troodontidae could be made on the loss can obscure even ordinal-level relationships basis of independent evidence. The morphology of neornithine taxa (Feduccia 1999, Paul 2002, of the braincase and otic region of troodontids is Murray and Vickers-Rich 2004, and references fundamentally avian and, in some respects, more therein). derived toward the condition in modern birds If the most birdlike maniraptorans were not than that of Archaeopteryx (e.g., hypoglossal [XII] flying and flightless radiations within Aves, we nerve with three or more external foramina in must regard their suites of flight-related char­ posterolateral region of braincase floor and con­ acters—such as those in the pectoral girdle and tralateral communication between at least some the presence of aerodynamic, asymmetrical flight tympanic diverticulae, among other characters; feathers in taxa like Microraptor—as exaptations see Currie 1985, Currie and Zhao 1993b). Troo­ (Paul 2002). Exaptational explanations (Gould dontid dentition and patterns of tooth implanta­ and Vrba 1982) for the origins of these characters tion are more like those of birds than like those of are common in the literature and are important nonmaniraptoran theropods (Currie 1987, Norell corollaries of the BMT hypothesis, because they and Hwang 2004). As noted above, the analysis of are inferred from the position of maniraptoran Mayr et al. (2005) is compatible with avian status and avian taxa on most cladograms (e.g., Padian for Troodontidae. and Chiappe 1998, Chiappe and Dyke 2002, Xu Less clear is whether the Alvarezsauridae rep­ 2006). Although some or all of the flight-related resent a flightless radiation within Aves. As noted characters of birds, including those found in the earlier, their phylogenetic relationships remain maniraptorans, may have evolved for purposes contentious, but some studies support their clas­ other than flight, current exaptational explana­ sification as birds more derived than Archaeop- tions offered by BMT proponents are often not teryx (e.g., Perle et al. 1993, 1994; Chiappe et al. fully formulated and rarely offer a biologically 1998; Padian 2004), and in our reanalysis of the plausible hypothesis to account for their origin CNM matrix, constraining alvarezsaurids to be (Feduccia 1985, 1993, 1995; Paul 2002; Feduccia within Aves made no difference in the length of et al. 2005, 2007). Under such conditions, exap­ MPTs. The morphology of the bizarre Therizino­ tational explanations should not be regarded sauroidea is much less like that of birds than is as having priority (Rose and Lauder 1996), and the morphology of oviraptorosaurs, dromaeo­ adaptational accounts should not be discarded. saurs, and troodontids. If the most birdlike maniraptoran clades belong Our results are compatible with these indepen­ within Aves, problematic exaptational expla­ dent data in suggesting avian status for at least nations, including those for the origin of flight dromaeosaurs and oviraptorosaurs and perhaps feathers, are unnecessary. for some other currently nonavian manirap­ The potential avian status of the most bird­ torans. Nevertheless, we appreciate the argu­ like maniraptoran taxa should not be dismissed ment that regarding clades of maniraptorans as simply because it does not appear in most clado­ flightless and flying radiations within Aves leads grams. As noted earlier, particularly worrisome

OM66_Text.indd 31 3/31/09 5:55:06 PM 32 ORNITHOLOGICAL MONOGRAPHS NO. 66

for cladistic evaluation are the characteristic re­ Aves—character support for the BMT hypothesis versals and convergences known to be associated would not be as overwhelming as is now claimed with flight loss (Feduccia 1999, 2002; Paul 2002). (Feduccia 2002, Feduccia et al. 2005). Although the As in modern birds, flight loss in Mesozoic birds BMT hypothesis and the modified theropod hy­ can involve severe reduction of the and pothesis of Paul (2002) would still be potentially pectoral girdle, can occur in both aquatic and supported, these hypotheses would be deprived terrestrial birds, and can involve gigantism. The of their most compelling character support. De­ Hesperornithes (Marsh 1880, Galton and Mar­ spite continuing attempts to reinterpret the mor­ tin 2002) are the best-known lineage of flightless phology of Archaeopteryx as that of an earthbound Mesozoic birds, but at least two other lineages predatory dinosaur (e.g., Paul 2002; Mayr et al. of ornithurines in the Mesozoic were flightless: 2005, 2007), comparisons of the morphologies the hen-sized terrestrial Patagopteryx (Chiappe of Archaeopteryx, the most basal avian, and non­ 2002b), which showed reduction of the forelimb maniraptoran theropods in general are not fa­ and pectoral girdle similar to that in modern vorable (Martin 1985, 1991, 1995; Feduccia 1999, flightless birds, and the large but poorly known 2002; Feduccia et al. 2005, 2007). Problems with Gargantuavis (Buffetaut and Le Loeuff 1998). Such homology statements, character transformations, phenomena all tend to worsen the problems and corollary issues like the origin of flight would posed by homoplasy in cladistic analyses and be magnified and simply shifted from birds and make deriving accurate inferences about phy­ maniraptorans onto birds and nonmaniraptoran logenetic relationships more difficult (Livezey theropods (Feduccia 2002, Feduccia et al. 2005). 1998, 2003). These ordinary problems posed by The principal characters potentially uniting Aves flight loss would be telescoped, in this case, to the and nonmaniraptoran theropods are the possible very base of avian evolution (Paul 2002). Mosaic presence of feathers, an advanced mesotarsus, character evolution in early avian history only and a functionally tridactyl foot. further complicates the situation (e.g., Zhou 2004, Unequivocally vaned feathers are restricted, Feduccia 2006). Also, the use of bipedal coeluro­ at present, to the maniraptoran clades Ovirap­ saurian outgroups, as in the analysis by Clark et torosauria and Dromaeosauridae. Fibrous struc­ al. (2002), may be contributing to a potentially tures in compsognathids (Currie and Chen 2001), misleading topology. Outgroup choice deter­ tyrannosauroids (Xu et al. 2004), alvarezsaurids mines the polarity of character states, including (Schweitzer et al. 1999, Schweitzer 2001), theriz­ ancestral reconstructions for entire clades (Nixon inosauroids (Xu et al. 1999a), and the dromaeosaur and Carpenter 1993). In this case, using bipedal Sinornithosaurus (Ji et al. 2001) cannot be unequiv­ cursors as outgroups may obscure phylogenetic ocally identified as feathers, despite widespread signal by wrongly treating characters indicating assertions to the contrary (e.g., Chiappe and flight loss as plesiomorphy. Dyke 2002, Norell and Xu 2005). These fibrous Weakened support for the BMT hypothesis and the structures are indistinguishable from decayed neoflightless-theropod hypothesis.—As noted, be­ dermal tissue, particularly degraded cause the two alternatives to the BMT hypothesis bundles (Feduccia 2002; Lingham-Soliar 2003a, b; supported by our results include the avian status Feduccia et al. 2005; Lingham-Soliar et al. 2007; of maniraptorans as a significant component, our Lingham-Soliar and Wesley-Smith 2008). Even if results and the independent evidence reviewed these fibrous structures are external integumen­ above support those alternatives. At the very tary appendages, their homology with avian least, our results supporting the avian status of feathers remains in doubt. Similar structures are some maniraptorans, which concur with inde­ preserved in ornithischian dinosaurs and in pte­ pendent evidence, weaken support for the BMT rosaurs (Unwin and Bakhurina 1994, Feduccia hypothesis. A representative sample of some of 2002, Ji and Yuan 2002, Wang et al. 2002, Ling­ the major putative synapomorphies of birds and ham-Soliar 2003a, Feduccia et al. 2005, Unwin theropods shows that many are restricted to the 2006), so these structures may not be homologous Maniraptora (see Table 6, where 21 of 29 char­ with avian feathers (see also Wellnhofer 2004). Ji acters listed are found only in maniraptorans). and Yuan (2002) and Czerkas and Ji (2002) re­ If some maniraptorans belong within Aves— garded the fibrous integumentary structures of and, hence, were removed from consideration as potentially homologous with avian in avian ancestry or as potential sister-groups of feathers, implying that feathers are basal to the

OM66_Text.indd 32 3/31/09 5:55:06 PM origin of birds 33

Table 6. Some of the major potential synapomorphies with birds that support the three major alternative hypotheses for the origin of birds, as suggested by our analysis or postulated in the literature. A “+” indicates that the character is present in at least one taxon, “−” indicates that the character is not present in any taxon, “?” indicates uncertainty as to whether the character is present in any taxa.

BMT Early-archosaur Crocodylomorph hypothesisa hypothesisb hypothesisc

Elongate integumentary structures wider at tips than at bases + d + e − Vaned integumentary structures + d + e − Accessory antorbital fenestra(e) in antorbital fossa + + − Orbits inflated with frontals strongly arched + d + − Postfrontal absent + + + Postorbital with ascending process of jugal reduced, descending − + − process of postorbital ventrally elongate Secondary quadrate articulation with braincase + d ? + Forward migration of quadrate head + d ? + Anterior tympanic recess + d ? + Dorsal tympanic recess + d ? + Posterior tympanic recess opening in columellar recess + d ? + Secondary tympanic membrane opening as fenestra pseudorotunda + d ? + because of lateral shift of perilymphatic duct Elongation of lagena and formation of tubular, elongate cochlear ? ? + f recess Vestibule in line with long-axis of cochlear recess ? ? + f Basioccipital prominently involved in formation of lower end of ? ? + f cup housing lagena Cerebral branches of internal carotid arteries describe sigmoid − ? + f curves along sides of cranium before meeting in pituitary fossa Pneumatic articular + g ? + Teeth lacking serrations, with constrictions between roots and + d + h + crowns, roots swollen Teeth covered by cementum and attached by periodontal ligament + d + h + Oval resorbtion pit perforating tooth lingually and sealed ventrally + d + h + Loss of interdental plates + d + h + Bones hollow + ? + Postcranial pneumaticity + g ? − Ossified uncinate processes + d − − Ossified sterna + − + Furcula + g + − Strap-like scapula parallel to vertebral column + d + − Scapulocoracoid at right angle rather than obtuse + d − − Elongation of coracoid into a strut-like element + d − + Coracoid with sternal grooves + d − + Lateral rotation around distal end of ulna when forelimb is flexed, + d ? + causing some degree of automatic folding of the manus Semilunate distal carpal + f − − Digits I, II, III prominent; IV and V reduced or lost + f − − Opisthopubic + ? + i Exclusion of from − ? + Ascending process of astragalus + ? − Advanced mesotarsus + ? − Foot functionally tridactyl + ? −

a BMT = birds are maniraptoran theropod dinosaurs. Sources are primarily Padian and Chiappe (1998), Sereno (1999), Holtz (2000), Norell et al. (2001), Zhou (2004), and Xu (2006). b Sources are primarily Jones et al. (2000, 2001), Martin (2004), and F. James and J. Pourtless (pers. obs.). c Sources are primarily Walker (1972, 1977, 1980, 1985, 1990), Whetstone and Martin (1979), Martin et al. (1980), Whetstone and Whybrow (1983), and Martin and Stewart (1999). d Only present in some maniraptoran taxa. e Character included in our matrix but not coded for Longisquama, for conservatism in treatment of contentious homologies. f Character not included in our matrix because of problems scoring the character for included taxa (see Methods). g Distribution uncertain. h Martin (2004), L. Martin (pers. comm.), F. James and J. Pourtless (pers. obs.). i Only known to be present in Hallopus.

OM66_Text.indd 33 3/31/09 5:55:07 PM 34 ORNITHOLOGICAL MONOGRAPHS NO. 66

clade stemming from the last common ancestor avian feathers, they may nevertheless not be ho­ of pterosaurs and birds, but no evidence from the mologous with modern avian feathers (see also fossil record indicates such a distribution. Sawyer and Knapp 2003). In light of these data, Some research on feather development has tests suggesting that fibrous structures preserved argued that feathers originated as evolutionarily with the alvarezsaurid Shuvuuia, structurally sim­ novel structures not homologous with reptil­ ilar to those found in other coelurosaurs, show ian scales and that the basal morphology of the β keratin–specific immunological reactivity and feather corresponds to the fibrous structures are, therefore, homologous with modern avian found in coelurosaurs like Sinosauropteryx (Prum feathers (Schweitzer et al. 1999, Schweitzer 2001, 1999, Prum and Brush 2002, Chuong et al. 2003, Paul 2002), should be interpreted cautiously. Prum and Dyck 2003). If true, this proposal would An advanced mesotarsus and functionally offer support to the identification of such struc­ tridactyl foot are compelling potential synapo­ tures as feathers. However, other developmental morphies with which to unite birds and non­ and molecular work supports the homology of maniraptoran theropods. Nevertheless, repeated feathers and reptilian scales and does not support trends toward bipedalism in Archosauria have the homology of feathers and coelurosaur fibrous long been recognized (e.g., by Romer 1956), all structures (e.g., Maderson and Alibardi 2000, Fe­ involving similar morphological alterations asso­ duccia 2002, Sawyer and Knapp 2003, Sawyer et ciated with the biophysical constraints of bipedal al. 2003a). Moreover, fossil evidence has revealed locomotion (e.g., Tarsitano 1991). Bipedalism is the presence of elongate tail feathers in several known to induce similar changes in unrelated taxa of Mesozoic birds (e.g., Protopteryx, Parapro- organisms (e.g., Romer 1956, Coombs 1978, Tarsi­ topteryx) that have been interpreted as scale-like, tano 1991, Berman et al. 2000). Trends toward bi­ again supporting the homology of feathers and pedalism are apparent even in nonarchosaurian reptilian scales (Zhang and Zhou 2000, Zheng et diapsids; the pelvis and hind limb of the al. 2007). Similar feathers are also present in the bolosaurid Eudibamus are modified for cursorial enigmatic (Zhang et al. 2008), the bipedal locomotion and parasagittal sister taxon to , which is probably posture, as in theropods (Berman et al. 2000). Sim­ not a theropod (Czerkas and Yuan 2002). Sawyer ilar trends toward bipedalism are evident in the et al. (2003b:27) observed that in turkeys (Melea- basal archosaur Euparkeria (Ewer 1965), in more gris), beard bristles, which are structurally similar derived crurotarsal taxa like Postosuchus and Orni- to the fibrous structures identified as feathers in thosuchus (Walker 1964, Chatterjee 1985, Feduccia coelurosaurs, display “simple branching, are hol­ 1999), and in the “sphenosuchian” crocodylo­ low, distally, and express the feather-type β kera­ morphs (Colbert 1952; Walker 1972, 1990; Crush tins,” even though they are not feathers. Sawyer 1984). In all cases where crurotarsal archosaurs et al. (2003b:30) argued that developed bipedalism, their skeletal , particularly in the pelvic girdle and hind limb, the present study raises the possibility that [the] “filamentous integumentary appendages” [of closely converged upon the anatomy of avemeta­ coelurosaurs] may more closely resemble the tarsalian archosaurs, especially theropods (Romer bristles of the wild beard, and may not 1956, Walker 1964, Chatterjee 1985, Feduccia 1999, depict intermediate stages in the evolution of Nesbitt 2007, and references therein). In the case feathers of Effigia, the trend toward bipedalism produced extreme convergence throughout the cranial and and concluded that postcranial skeleton on that of highly cursorial Without more detailed information about the cel­ ornithomimosaurs, with further convergence on lular and molecular nature of the “filamentous numerous characters of the avemetatarsalian, integumentary appendages” of non-avian dino­ dinosaurian, theropod, neotetanurine, and coelu­ saurs, and more information on the presence or rosaurian skeletons; in some cases, the relevant absence of follicles, it may be premature to as­ characters are identical across taxa (Nesbitt and sume the homology of all these “filamentous in­ Norell 2006, Nesbitt 2007). Nesbitt et al. (2007:71) tegumentary appendages” with feathers. concluded that the “many convergences be­ Therefore, even if the fibrous structures of coelu­ tween Effigia and ornithomimids suggest that a rosaurs are external integumentary appendages ‘theropod-like ’ developed in a group that are structurally and chemically similar to of -line archosaurs before it evolved in

OM66_Text.indd 34 3/31/09 5:55:07 PM origin of birds 35

later theropod dinosaurs.” Clearly, the conver­ distributed among maniraptorans. Clearly, then, gence upon dinosaurian and avemetatarsalian this character does not strongly support mono­ anatomy seen in numerous lineages of crurotar­ phyly of Theropoda as presently constituted. sal archosaurs is more than superficial. Although Interestingly, analysis of our matrix supported none of these taxa possesses an advanced me­ the possibility of monophyly of a weakly sup­ sotarsus, a simple mesotarsus is present in Eu- ported clade of theropods excluding manirap­ parkeria. Walker (1972, 1977) outlined a plausible torans (Figs. 9–10). If some maniraptorans were functional scenario whereby an advanced meso­ birds, and if birds were not theropods, similari­ tarsus could be derived from a crurotarsal joint ties between maniraptorans and theropods could as trends toward bipedalism continued (see also be readily explained by convergence on a curso­ Romer 1956). Moreover, in facultatively bipedal rial morphotype subsequent to the loss of flight. crurotarsal taxa, the outer metatarsals decrease Even distantly related reptiles could converge in length and the metatarsus becomes more elon­ closely, in some cases almost indistinguishably, gate and compact, approximating the condition on the theropod morphotype through the acqui­ in theropods, although in none of these taxa is sition of cursoriality, as the case of Effigia, noted the foot functionally tridactyl (e.g., Colbert 1952, above, dramatically demonstrates (Nesbitt and Crush 1984, Walker 1990). As noted by Walker Norell 2006, Nesbitt 2007). If this is the case, some (1977), Martin (1983, 2004), and Feduccia (1999), maniraptorans represent lineages of cryptic birds the biophysical pressures operating on an arbo­ whose true phylogenetic relationships have been real organism leaping between trees are similar obscured by convergence and the loss of flight. to those operating on a terrestrial biped, which Given the evidence that some maniraptorans suggests that the advanced mesotarsus and func­ may belong within Aves and that, consequently, tionally tridactyl foot of birds may have arisen by Aves may not belong within Theropoda, this pos­ a different functional trajectory from that of non­ sibility must be seriously considered. maniraptoran theropods. Note that (1) if the core maniraptorans, at least, These three characters (the possible presence are not dinosaurs but rather a cryptic lineage of feathers, an advanced mesotarsus, and a func­ of birds, and (2) if Aves does not belong within tionally tridactyl foot) do not unambiguously Theropoda, as is likely if core maniraptorans are support the alignment of Aves (inclusive of some in fact birds and (3) if the inference of a manual maniraptorans) and nonmaniraptoran theropods digital identity of II, III, and IV in modern birds to the exclusion of other archosaurian taxa. Given is correct (e.g., Burke and Feduccia 1997, Feduc­ these data, and our results supporting alterna­ cia and Nowicki 2002, Kundrát et al. 2002), then tives to the BMT and neoflightless-theropod hy­ the manus of core maniraptorans is probably potheses, Theropoda as presently constituted composed of digits II, III, and IV. If this inference may not be monophyletic. Aves, including at least is correct, determination of carpal and manual some maniraptorans, may be more closely related homologies between birds and maniraptorans to some other archosaurian taxon. The claim would be simplified (see Appendix 3), but neither that some maniraptorans may be birds and not the homology of manual digits of core manirap­ theropods is not contradictory. If some manirap­ torans and birds nor the homology of manual torans belong within Aves, their phylogenetic re­ digits of theropods in general and birds can be lationships with theropods are contingent upon determined at present, and to assume that they whether Aves belongs within Theropoda, and, as are homologous is unwarranted. noted, comparisons of birds and those theropods Are alternatives to the BMT hypothesis compatible that remain when Maniraptora are reclassified with the evidence?—Analysis of our new matrix as birds are not favorable. Note that, although indicated that both the early-archosaur and cro­ Theropoda is widely accepted as a well-supported codylomorph hypotheses are at least as strongly clade, the analysis of Benton (2004) demonstrated supported as the BMT. In addition to the support that the only unequivocal synapomorphy diag­ for these hypotheses derivable from our trees, nosing Theropoda is the presence of an intraman­ we found, using Kishino-Hasegawa tests, that dibular joint. An intramandibular joint is also the topologies of the crocodylomorph and early- present in crurotarsal taxa like Postosuchus (Chat­ archosaur hypotheses were at least as likely as the terjee 1985) and Ornithosuchus (Walker 1964), is standard BMT hypothesis (the hypothesis of no not primitively present in birds, and is erratically difference could not be rejected). P values for the

OM66_Text.indd 35 3/31/09 5:55:08 PM 36 ORNITHOLOGICAL MONOGRAPHS NO. 66

other hypotheses were smaller, which suggests and Paraprotopteryx (Zheng et al. 2007) are some­ larger differences (Table 3). The hypothesis that what similar, at least in overall appearance, to the birds evolved twice and Paul’s hypothesis that integumentary structures of Longisquama. Our some maniraptorans belong within Aves but that recovery of Longisquama in a weakly supported Aves still belongs within the Theropoda also ap­ clade with birds and maniraptorans (Figs. 9 and peared to be less likely in terms of the number of 12) lends some support to the interpretation of steps in their MPTs. The alternative hypothesis of Jones et al. (2000, 2001) and Martin (2004). Kurochkin (2006a, b), postulating avian diphyly, Unfortunately, a detailed osteology of Long- is least supported. Also unclear is what charac­ isquama has not been published. Martin (2004) has ters could unite the derived ornithurines (which identified potential synapomorphies shared by are already morphologically advanced when they the cranial and pectoral skeletons of Longisquama appear in the fossil record; e.g., Feduccia 1999) and birds that have previously been overlooked, with an unspecified archosauromorph lineage. including a subdivided antorbital fenestra (Mar­ In addition to the results obtained through use tin 2004, fig. 4; Fig. 15), which other authors have of Kishino-Hasegawa tests, we also recovered a regarded as an important synapomorphy of clade of maniraptorans, birds, and the basal ar­ birds and theropods (e.g., Paul 2002). Examina­ chosaur Longisquama, though we note that it was tion of the skull of Longisquama as preserved on only weakly supported (Figs. 9 and 12). These the counterslab of the main specimen supports results nevertheless support the possibility of a Martin’s interpretation (Fig. 15). A preorbital va­ sister-group relationship between Longisquama cuity delimited by the lacrimal posteriorly, the and Aves (inclusive of some maniraptorans). In ventrally, and the nasal dorsally is a large addition, birds and maniraptorans were never antorbital fenestra; the rostral portion of the skull unambiguously associated with nonmanirap­ beyond the antorbital fossa is not preserved (the toran theropods in any of our trees (Figs. 9–13). slab is broken just beyond the antorbital fossa). A variant of the crocodylomorph hypothesis may The size and location of this vacuity, extending also be correct, because the polytomies recovered from the rostral portion of the skull posteriorly to in Figures 10 and 13 do not preclude at least a the anterior border of the lacrimal, is inconsistent sister-group relationship between birds and with identification as the naris. It can only be the crocodylomorphs, agreeing with our Kishino- antorbital fenestra. The ventral swelling of the na­ Hasegawa tests. sal and the dorsal swelling of the maxilla are simi­ Our results with respect to the early-archosaur lar to the processes observed in other archosaur hypothesis concur with some independent evi­ taxa with subdivided antorbital fenestrae and dence. Diagnosable potential synapomorphies support the inference that an accessory antorbital exist with which to unite Longisquama and Aves fenestra was present. In fact, the lower half of the (inclusive of some maniraptorans; Table 6). To be strut of bone arising from these processes and de­ conservative, we did not score the integumen­ limiting the posterior edge of the accessory antor­ tary structures associated with the type skel­ bital fenestra is visible above the dorsal swelling eton of Longisquama as structurally homologous of the maxilla. The orbit is considerably inflated, with avian feathers. The data provided by Jones causing moderate vaulting of the frontals. The et al. (2000 [figs. 2–6], 2001) and Martin (2004) posterior skull is enlarged, and the parietals are support the interpretation of the integumentary bulged, which suggests expansion of the brain. structures as homologous with avian feathers, Senter (2004) argued that the parietals were orna­ but Prum (2001), Unwin and Benton (2001), and mented with ridges or bumps, but these appear to Paul (2002), among others, have disagreed. Voigt be a preservation defect of the posterior skull (F. et al. (2009) did not consider the integumentary James and J. Pourtless pers. obs.). The temporal structures of Longisquama to be homologous with fenestrae appear to be confluent with the orbit, as either avian feathers or reptilian scales. Never­ in birds; alternatively, the upper temporal fenestra theless, they concluded that the integument of may be preserved (question mark in Fig 15; but Longisquama shares several morphological char­ this feature is probably a preservation defect). The acters with feathers and developed through a long postorbital was dislodged postmortem and two-stage process similar to that by which feath­ now intrudes into the orbit; its base is anterior to ers develop. Interestingly, the long scale-like the small ascending process of the jugal. Conse­ feathers of Protopteryx (Zhang and Zhou 2000) quently, the orbit would have been rounder in

OM66_Text.indd 36 3/31/09 5:55:08 PM origin of birds 37

Fig. 15. Skull of Longisquama insignis as preserved on the counterslab of the type specimen (left) and a recon­ struction and interpretation of the cranial osteology (right). Note the birdlike overall appearance and construction of some of the cranial elements. See text (Discussion section) for details. Abbreviations: af = antorbital fossa, aof = antorbital fenestra, d? = dentary and other mandibular elements, dpm = dorsal process of maxilla supporting the strut of bone subdividing the antorbital fenestra and delimiting an accessory antorbital fenestra, f = frontal, j = jugal, l = lacrimal, n = nasal, o = orbit, par = paroccipital, po = postorbital, sr = supraorbital ridge of the frontal, vp = ventral process of the nasal supporting the strut of bone subdividing the antorbital fenestra and delimiting an accessory antorbital fenestra, and ? = a preservation defect or, less probably, the upper temporal fenestra. Dashes indicate breaks in the bone or in the matrix. Dotted lines indicate inferred or partially preserved structures or uncertain borders of cranial elements. Photograph by John Ruben, provided by courtesy of Larry Martin.

. Interestingly, Longsiquama has the same den­ Aves (inclusive of some maniraptorans; Table tal morphology style as birds and crocodyliform 6), either as sister taxa, as proposed by Walker crocodylomorphs (Martin 2004, L. Martin pers. (1972), or perhaps according to the topologies comm., F. James and J. Pourtless pers. obs.; Fig. 15). proposed by Martin et al. (1980), Whetstone and The dentary and mandibular elements are not Whybrow (1983), and Martin and Stewart (1999). well preserved on the counterslab. The appen­ The structures of the skull, and in particular of the dicular skeleton is generalized, but a furcula is braincases and otic regions of crocodylomorphs present, and the scapula is straplike (Sharov 1970; and birds, show numerous detailed similarities, Jones et al. 2000 [fig. 1], 2001). Unfortunately, the including extensive pneumaticity associated with pelvic girdle and hind limb of Longisquama are not the ear (Walker 1972 [figs. 1, 5, 6], 1990 [e.g., figs. known, so further comparisons with birds are not 15–30, 48, 51, 52]; Whetstone and Martin 1979). currently possible. The subdivision of the antor­ The dentitions and dental systems of crocodyli­ bital fenestra, the expansion of the orbit, the mor­ form crocodylomorphs and toothed birds are phology of the postorbital, the expansion of the nearly indistinguishable (Martin et al. 1980 [fig. posterior skull suggesting expansion of the brain, 2]; Martin 1985 [fig. 1], 1991 [figs. 15–18];­ Mar the thin jugal with small ascending process, the tin and Stewart 1999 [figs. 1–3]). Some evidence dental morphology, and the presence of a furcula also implies that some crocodylomorphs, like provide osteological support for an association the “sphenosuchians” (on paraphyly of “Sphe­ of Longisquama with birds independently of the nosuchia,” see Clark and Sues 2002, Sues et al. integumentary structures. We do not assert that 2003), were descended from arboreal ancestors Longisquama is the sister taxon of Aves (inclusive (Walker 1972). An arboreal origin of flight could of some maniraptorans), but such a relationship therefore be compatible with a crocodylomorph is one possibility that is supported by our results hypothesis for the origin of birds (Walker 1972). and by some character evidence. Some crocodylomorphs had long, ossified sterna Some independent evidence also supports our (Crush 1984, fig. 7B). The coracoids of “sphenosu­ results with respect to the crocodylomorph hy­ chians” and birds are not dissimilar, and the latter pothesis. Several diagnosable potential synapo­ could easily be derived from the former (Walker morphies could unite crocodylomorph taxa and 1972, fig. 7). During flexion of the in

OM66_Text.indd 37 3/31/09 5:55:09 PM 38 ORNITHOLOGICAL MONOGRAPHS NO. 66

crocodylomorphs, mechanical linkage of ele­ Wilkinson (1996), that phylogenetic relationships ments causes the manus and carpus to rotate lat­ among Archosauria are poorly understood (e.g., erally around the distal end of the ulna, which is Romer 1956, Charig 1993, Feduccia 1999). Clearly, similar to avian wing folding (Walker 1972, fig. extensive homoplasy is at work. In such cases, 8). At least one crocodylomorph, Hallopus, may even rigorous applications of cladistic methods even have had an opisthopubic pelvis (Walker can fail to resolve phylogenetic relationships ac­ 1970, 1972 [fig. 9]). Some characters regularly of­ curately (e.g., Gosliner and Ghiselin 1984, Car­ fered as indicating a close relationship between roll 1988, Carroll and Dong 1991, Wiens et al. birds and theropods, such as a light skeleton with 2003). The present uncertainties in discriminating hollow bones, are also found in “sphenosuchian” among the BMT, early-archosaur, and crocodyl­ crocodylomorphs (Colbert 1952; Walker 1972, omorph hypotheses are an invitation to further 1990; Crush 1984; Table 6). An obvious problem study. The neoflightless-theropod hypothesis of with this hypothesis is the conversion of the cru­ Paul, or even the “birds-evolved-twice” hypoth­ rotarsal ankle to a mesotarsal ankle, but, as noted esis of Kurochkin, may prove correct in the light earlier, Walker’s (1972, 1977) solution to this of future discoveries. Such uncertainty drives the problem merits further attention. Again, we do growth of scientific knowledge, and it should be not assert that a variant of the crocodylomorph welcomed rather than discounted. hypothesis is correct, but this is another possibil­ ity that is supported by our results and by some Conclusions character evidence. Despite the support our analysis offers to these We have pursued two goals: evaluation of two alternatives to the BMT hypothesis, and in whether the BMT hypothesis is as well supported spite of the failure of our analyses to support pre­ as has been claimed, and evaluation of alternative dictions derivable from the BMT, we reiterate that hypotheses for the origin of birds within a com­ our analyses do not permit a single hypothesis for parative phylogenetic framework. We conclude the origin of birds to be singled out as correct at that, because of circularity in the construction this time. Our results neither conclusively refute of matrices, inadequate taxon sampling, insuffi­ the BMT hypothesis nor conclusively support ciently rigorous application of cladistic methods, either of the two alternatives discussed above. and a verificationist approach, the BMT hypoth­ However, our analysis indicates (1) that, of the esis has not been subjected to sufficiently rigor­ five current hypotheses for the origin of birds, the ous attempts at refutation, and the literature does BMT hypothesis, the early-archosaur hypothesis, not provide the claimed overwhelming support. and a variant of the crocodylomorph hypoth­ Our analyses and independent data indicate that esis are most compatible with presently avail­ two of the alternatives to the BMT hypothesis are able data; and (2) that the BMT hypothesis is not as probable as the BMT and are potentially sup­ clearly preferable to these two alternatives. Dis­ ported by specific osteological data. These alterna­ crimination among these three hypotheses may tives are the early-archosaur hypothesis, positing be aided by derivation of predictions from each a sister-group relationship between Longisquama and performance of tests that establish whether and Aves, and a variant of the crocodylomorph these predictions are supported, but, at present, hypothesis. Both hypotheses include the proposi­ ambiguities in the data complicate discrimina­ tion that some maniraptorans are actually birds tion among the three. Insufficient information more derived than Archaeopteryx. from the fossil record (e.g., Heckert and Lucas Ostrom (1975, 1976a, b) and subsequent re­ 2003) and uncertainties about the homologies of searchers like Gauthier (1986) were correct key characters in archosaur evolution contribute in noting the extensive similarities between to this difficulty (see Appendix 3). The potential maniraptorans and birds, a conclusion only synapomorphies that may underlie any of these strengthened by more recent discoveries, but three topologies also overlap considerably (Ta­ evidence suggests that at least some manirap­ ble 6). These data and our results, consistently torans belong within Aves. If Aves (inclusive of retrieving a polytomy of archosaurs more de­ some maniraptorans) does not belong within rived than basal-most forms like Euparkeria and Theropoda, at least some maniraptorans should Erythrosuchus (Figs. 9–13), support the earlier be classified as birds rather than dinosaurs, and consensus, partially challenged by Gower and Aves should not be considered a lineage of living

OM66_Text.indd 38 3/31/09 5:55:09 PM origin of birds 39

dinosaurs. On the basis of our results, the next Anatomica Avium, 2nd ed. (J. J. Baumel, A. S. King, two major challenges are to evaluate further the J. E. Breazile, H. E. Evans, and J. C. Vanden Berge, possibility that some maniraptorans in fact be­ Eds.). Nuttall Ornithological Club, Cambridge, long within Aves, rather than the reverse, and to Massachusetts. Beebe, C. W. 1915. A tetrapteryx stage in the ancestry of further explore whether birds may have been de­ birds. Zoologica 2:38–52. rived from theropods, “early archosaurs,” or cro­ Bellairs, A. d’A., and C. R. Jenkin. 1960. The skeleton codylomorphs, the three most likely candidates of birds. Pages 241–300 in Biology and Compara­ given current evidence. At present, the origin of tive Physiology of Birds, vol. 1 (A. J. Marshall, Ed.). birds is an open question. Academic Press, New York. Bennett, S. C. 2008. Ontogeny and Archaeopteryx. Jour­ Acknowledgments nal of Vertebrate Paleontology 28:535–542. Benton, M. J. 1999. Scleromochlus taylori and the origin of dinosaurs and pterosaurs. Philosophical Trans­ We are especially grateful to M. Holder, who guided actions of the Royal Society of London, Series B us through the construction of our matrix and our 354:1423–1446. entire analysis and spent many hours discussing sys­ Benton, M. J. 2004. Origin and relationships of Dino­ tematics with us. For general advice and comments on sauria. Pages 7–19 in The Dinosauria, 2nd ed. (D. B. earlier drafts of the manuscript, we thank P. Beerli, D. Weishampel, P. Dodson, and H. Osmólska, Eds.). Burnham, F. Crews, P. Currie, G. Dyke, A. Feduccia, A. University of California Press, Berkeley. James, H. James, P. Kareiva, E. Kurochkin, L. Martin, G. Benton, M. J., and A. D. Walker. 2002. Erpetosuchus, Mayr, C. McCulloch, G. Naylor, S. Olson, G. Paul, D. a crocodile-like basal archosaur from the Late Tri­ Steadman, S. Steppan, D. Swofford, K. van der Linde, J. assic of Elgin, Scotland. Zoological Journal of the Wilgenbusch, and R. Zusi. For help with the literature Linnean Society 136:25–47. and access to collections we thank G. Erickson (Florida Berman, D. S., R. R. Reisz, D. Scott, A. C. Henrici, State University), L. Martin (University of Kansas), S. S. Sumida, and T. Martens. 2000. Early Permian and P. Currie and J. Gardner (Royal Tyrell Museum). bipedal . Science 290:969–972. We thank K. Womble and S. Bonner for constructing Brochu, C. A. 2003. Osteology of Tyrannosaurus rex: our figures. For outstanding editorial assistance, we Insights from a Nearly Complete Skeleton and thank A. B. Thistle. For Russian translations, we thank High-resolution Computed Tomographic Analy­ M. Ustymenko and D. Jeltovski. For information on sis of the Skull. Society of Vertebrate Paleontology the Theropod Working Group, see research.amnh.org/ Memoirs, no. 7. users/norell/TWGhome.html. Broom, R. 1913. On the South African pseudosuchian Euparkeria and allied genera. Proceedings of the Literature Cited Zoological Society of London 1913:619–633. Bryant, H. N., and A. P. Russell. 1993. The occurrence Barsbold, R., P. J. Currie, N. P. Myhrvold, H. Osmól- of within Dinosauria: Implications for the ska, K. Tsogtbaatar, and M. Watabe. 2000. A homology of the avian furcula and the utility of from a non-avian theropod. Nature negative evidence. Journal of Vertebrate Paleontol­ 403:155–156. ogy 13:171–184. Barsbold, R., and T. Marya ´nska. 1990. Segnosauria. Buffetaut, E., and J. Le Loeuff. 1998. A new giant Pages 408–415 in The Dinosauria (D. B. Weisham­ ground bird from the Upper Cretaceous of south­ pel, P. Dodson, and H. Osmólska, Eds.). University ern France. Journal of the Geological Society of of California Press, Berkeley. London 155:1–4. Barsbold, R., and H. Osmólska. 1990. Ornithomi­ Burgers, P., and L. M. Chiappe. 1999. The wing of mosauria. Pages 225–244 in The Dinosauria (D. B. Archaeopteryx as a primary thrust generator. Nature Weishampel, P. Dodson, and H. Osmólska, Eds.). 399:60–62. University of California Press, Berkeley. Burgers, P., and K. Padian. 2001. Why thrust and Barsbold, R., and H. Osmólska. 1999. The skull of ground effect are more important than lift in the Velociraptor (Theropoda) from the evolution of sustained flight. Pages 351–361in New of Mongolia. Acta Palaeontologica Polonica 44:189– Perspectives on the Origin and Early Evolution 219. of Birds: Proceedings of the International Sympo­ Barsbold, R., and A. Perle. 1984. On the first new sium in Honor of John H. Ostrom (J. Gauthier and find of a primitive ornithomimosaur from the Cre­ L. F. Gall, Eds.). Press, New Haven, taceous of the M.P.R. [In Russian.] Paleontological Connecticut. Journal 1984:121–123. Burke, A. C., and A. Feduccia. 1997. Developmental Baumel, J. J., and L. J. Witmer. 1993. Osteologia. Pages patterns and the identification of homologies in the 45–132 in Handbook of Avian Anatomy: Nomina avian hand. Science 278:666–669.

OM66_Text.indd 39 3/31/09 5:55:10 PM 40 ORNITHOLOGICAL MONOGRAPHS NO. 66

Burnham, D. A. 2004. New information on Bambiraptor Chiappe, L. M. 2002a. Basal bird phylogeny: Problems and feinbergi (Theropoda: Dromaeosauridae) from the solutions. Pages 448–472 in Mesozoic Birds: Above the Late Cretaceous of . Pages 67–111 in Feath­ Heads of Dinosaurs (L. M. Chiappe and L. M. Witmer, ered Dragons: Studies on the Transition from Dino­ Eds.). University of California Press, Berkeley. saurs to Birds (P. J. Currie, E. B. Koppelhus, M. A. Chiappe, L. M. 2002b. Osteology of the flightless Shugar, and J. L. Wright, Eds.). Indiana University Patagopteryx deferrariisi from the Late Cretaceous Press, Bloomington. of Patagonia (Argentina). Pages 281–316 in Meso­ Burnham, D. A. 2007. Paleoenvironment, paleoecology, zoic Birds: Above the Heads of Dinosaurs (L. M. and evolution of maniraptoran “dinosaurs.” Ph.D. Chiappe and L. M. Witmer, Eds.). University of dissertation, University of Kansas, Lawrence. California Press, Berkeley. Burnham, D. A., K. L. Derstler, P. J. Currie, R. T. Bak- Chiappe, L. M. 2007. Glorified Dinosaurs: The Origin ker, Z. Zhou, and J. H. Ostrom. 2000. Remarkable and Early . Wiley, Hoboken, New new birdlike dinosaur (Theropoda: Maniraptora) Jersey. from the Upper Cretaceous of Montana. University Chiappe, L. M., and G. J. Dyke. 2002. The Mesozoic of Kansas Paleontological Contributions 13:1–14. radiation of birds. Annual Review of Ecology and Carpenter, K., C. Miles, J. H. Ostrom, and K. Systematics 33:91–124. Cloward. 2005. Redescription of the small Chiappe, L. M., S. Ji, Q. Ji, and M. A. Norell. 1999. maniraptoran theropods Ornitholestes and Coelurus Anatomy and systematics of the Confuciusorni­ from the Upper of thidae (Theropoda: Aves) from the late Mesozoic Wyoming. Pages 49–71 in The Carnivorous Dino­ of northeastern China. Bulletin of the American saurs (K. Carpenter, Ed.). Indiana University Press, Museum of Natural History 242:1–89. Bloomington. Chiappe, L. M., M. A. Norell, and J. M. Clark. 1998. Carroll, R. L. 1988. Vertebrate paleontology and evo­ The skull of a relative of the stem-group bird lution. W. H. Freeman, New York. Mononykus. Nature 392:275–278. Carroll, R. L., and Z. M. Dong. 1991. Hupehsuchus, an Chiappe, L. M., M. A. Norell, and J. M. Clark. 2002. enigmatic aquatic reptile from the Triassic of China, The Cretaceous, short-armed Alvarezsauridae: and the problem of establishing relationships. Phil­ Mononykus and its kin. Pages 87–120 in Meso­ osophical Transactions of the Royal Society of Lon­ zoic Birds: Above the Heads of Dinosaurs (L. M. don, Series B 331:131–153. Chiappe and L. M. Witmer, Eds.). University of Charig, A. J. 1993. Recently proposed phylogenetic California Press, Berkeley. analyses of the Triassic Archosauria: A critical Chiappe, L. M., and C. Walker. 2002. Skeletal morphol­ comparison and evaluation, facilitated by a simple ogy and systematics of the Cretaceous Euenantio­ technique for the modification of conflicting den­ rnithines (: Enantiornithes). Pages drograms. Paleontologia Lombarda della Societa 240–267 in Mesozoic Birds: Above the Heads of Italiana di Scienze Naturali e del Museo Civico di Dinosaurs (L. M. Chiappe and L. M. Witmer, Eds.). Storia Natrale di Milano, Nuova Serie 2:45–62. University of California Press, Berkeley. Chatterjee, S. 1985. Postosuchus, a new thecodontian Chiappe, L. M., and L. M. Witmer, Eds. 2002. Mesozoic reptile from the Triassic of and the origin of Birds: Above the Heads of Dinosaurs. University of tyrannosaurs. Philosophical Transactions of the California Press, Berkeley. Royal Society of London, Series B 309:395–460. Christiansen, P. 2006. Allometry in phylogeny and Chatterjee, S. 1991. Cranial anatomy and relationships Archaeopteryx. Journal of Vertebrate Paleontology of a new Triassic bird from Texas. Philosophical 26:480–486. Transactions of the Royal Society of London, Series Chuong, C. M., P. Wu, F. Zhang, X. Xu, M. Yu, R. B. B 332:277–346. Widelitz, T. Jiang, and L. Hou. 2003. Chatterjee, S. 1997. The Rise of Birds: 225 Million to the sky: Defining the feather with integument of Evolution. Johns Hopkins University Press, Bal­ from Mesozoic China and experimental timore, Maryland. evidence from molecular laboratories. Journal of Chatterjee, S., and R. J. Templin. 2004. Feathered coe­ Experimental B 298:42–56. lurosaurs from China: New light on the arboreal Chure, D. J. 2001. The wrist of Allosaurus (Saurischia: . Pages 251–281 in Feathered Theropoda), with observations on the carpus in Dragons: Studies on the Transition from Dinosaurs theropods. Pages 283–300 in New Perspectives on to Birds (P. J. Currie, E. B. Koppelhus, M. A. Shugar, the Origin and Early Evolution of Birds: Proceed­ and J. L. Wright, Eds.). Indiana University Press, ings of the International Symposium in Honor of Bloomington. John H. Ostrom (J. Gauthier and L. F. Gall, Eds.). Chiappe, L. M. 1992. Enantiornithine (Aves) tarsometa­ Yale University Press, New Haven, Connecticut. tarsi and the avian affinity of the Late Cretaceous Chure, D. J., and J. H. Madsen, Jr. 1996. On the presence . Journal of Vertebrate Paleontology of furculae in some non-maniraptoran theropods. 12:344–350. Journal of Vertebrate Paleontology 18:573–577.

OM66_Text.indd 40 3/31/09 5:55:10 PM origin of birds 41

Clark, J. M., T. Marya ´nska, and R. Barsbold. 2004. Corfe, I. J., and R. J. Butler. 2006. Comment on “A Therizinosauroidea. Pages 151–164 in The Dino­ well-preserved Archaeopteryx specimen with thero­ sauria, 2nd ed. (D. B. Weishampel, P. Dodson, and pod features.” Science 313:1238b. H. Osmólska, Eds.). University of California Press, Cranston, K. A., and B. Rannala. 2007. Summarizing Berkeley. a posterior distribution of trees using agreement Clark, J. M., M. A. Norell, and R. Barsbold. 2001. subtrees. Systematic Biology 56:578–590. Two new oviraptorids (Theropoda: Oviraptorosau­ Cruickshank, A. R. I. 1972. The proterosuchian theco­ ria) Late Cretaceous Djadoktha Formation, Ukhaa donts. Pages 89–119 in Studies in Vertebrate Evo­ Tolgod, Mongolia. Journal of Vertebrate Paleontol­ lution: Essays Presented to F. R. Parrington (K. A. ogy 21:209–213. Joysey and T. S. Kemp, Eds.). Oliver and Boyd, Clark, J. M., M. A. Norell, and L. M. Chiappe. 1999. Edinburgh, . An oviraptorid skeleton from the Late Cretaceous Crush, P. J. 1984. A late Upper Triassic sphenosuchid of Ukhaa Tolgod, Mongolia, preserved in an avian- crocodilian from Wales. Palaeontology 27:131–157. like brooding position over an oviraptorid nest. Currie, P. J. 1985. Cranial anatomy of American Museum Novitates 3265:1–36. inequalis (Saurischia, Theropoda) and its bearing on Clark, J. M., M. A. Norell, and P. J. Makovicky. the origin of birds. Canadian Journal of Earth Sci­ 2002. Cladistic approaches to the relationships of ence 22:1643–1658. birds to other theropod dinosaurs. Pages 31–60 in Currie, P. J. 1987. Bird-like characteristics of the jaws Mesozoic Birds: Above the Heads of Dinosaurs and teeth of troodontid theropods (Dinosauria, (L. M. Chiappe and L. M. Witmer, Eds.). University Saurischia). Journal of Vertebrate Paleontology of California Press, Berkeley. 7:72–81. Clark, J. M., A. Perle, and M. Norell. 1994. The skull Currie, P. J. 1995. New information on the anatomy and of Erlicosaurus andrewsi, a Late Cretaceous “segno­ relationships of Dromaeosaurus albertensis (Dinosau­ saur” (Theropoda: Therizinosauroidea) from Mon­ ria: Theropoda). Journal of Vertebrate Paleontology golia. American Museum Novitates 3115:1–39. 15:576–591. Clark, J. M., and H.-D. Sues. 2002. Two new basal cro­ Currie, P. J., and P. Chen. 2001. Anatomy of Sinosau- codylomorph archosaurs from the Lower Jurassic ropteryx prima from , northeastern China. and the monophyly of the Sphenosuchia. Zoologi­ Canadian Journal of Earth Sciences 38:1705–1727. cal Journal of the Linnean Society 136:77–95. Currie, P. J., and Z. Dong. 2001. New information on Clark, J. M., H.-D. Sues, and D. S. Berman. 2000. A Cretaceous troodontids (Dinosauria, Theropoda) new specimen of Hesperosuchus agilis from the from the People’s Republic of China. Canadian Upper Triassic of New Mexico and the interrela­ Journal of Earth Sciences 38:1753–1766. tionships of basal crocodylomorph archosaurs. Currie, P. J., E. B. Koppelhus, M. A. Shugar, and Journal of Vertebrate Paleontology 20:683–704. J. L. Wright, Eds. 2004. Feathered Dragons: Studies Clark, J. M., J. Welman, J. A. Gauthier, and J. M. on the Transition from Dinosaurs to Birds. Indiana Parrish. 1993. The laterosphenoid bone of early University Press, Bloomington. archosauriforms. Journal of Vertebrate Paleontol­ Currie, P. J., and X. Zhao. 1993a. A new carnosaur ogy 13:48–57. (Dinosauria, Theropoda) from the Jurassic of Xinji­ Clarke, J. A. 2004. Morphology, phylogenetic taxon­ ang, People’s Republic of China. Canadian Journal omy, and systematics of Ichthyornis and of Earth Sciences 30:2037–2081. (Avialae: Ornithurae). Bulletin of the American Currie, P. J., and X. Zhao. 1993b. A new troodon­ Museum of Natural History 286:1–179. tid (Dinosauria, Theropoda) braincase from the Clarke, J. A., and M. A. Norell. 2002. The morphol­ () of Alberta. ogy and phylogenetic position of Apsaravis ukhaana Canadian Journal of Earth Sciences 30:2231–2247. from the Late Cretaceous of Mongolia. American Czerkas, S. A., and Q. Ji. 2002. A new rhamphorhyn­ Museum Novitates 3387:1–46. choid with a headcrest and complex integumentary Colbert, E. H. 1952. A pseudosuchian reptile from Ari­ structures. Pages 15–41 in Feathered Dinosaurs and zona. Bulletin of the American Museum of Natural the Origin of Flight (S. J. Czerkas, Ed.). Dinosaur History 99:565–592. Museum Journal, vol. 1. Dinosaur Museum, Bland­ Colbert, E. H. 1989. The Triassic dinosaur Coelophy- ing, Utah. sis. Bulletin of the Museum of Northern Arizona Czerkas, S. A., and C. Yuan. 2002. An arboreal 57:1–160. maniraptoran from northeast China. Pages 63–95 Colbert, E. H., and D. A. Russell. 1969. The small in Feathered Dinosaurs and the Origin of Flight Cretaceous dinosaur Dromaeosaurus. American (S. J. Czerkas, Ed.). Dinosaur Museum Journal, Museum Novitates 2380:1–49. vol. 1. Dinosaur Museum, Blanding, Utah. Coombs, W. P., Jr. 1978. Theoretical aspects of curso­ Czerkas, S. A., D. Zhang, J. Li, and Y. Li. 2002. Fly­ rial in dinosaurs. Quarterly Review of ing dromaeosaurs. Pages 97–126 in Feathered Biology 53:393–418. Dinosaurs and the Origin of Flight (S. J. Czerkas,

OM66_Text.indd 41 3/31/09 5:55:11 PM 42 ORNITHOLOGICAL MONOGRAPHS NO. 66

Ed.). Dinosaur Museum Journal, vol. 1. Dinosaur Feduccia, A. 1999. The Origin and Evolution of Birds, 2nd Museum, Blanding, Utah. ed. Yale University Press, New Haven, Connecticut. Dahn, R. D., and J. F. Fallon. 2000. Interdigital regula­ Feduccia, A. 2002. Birds are dinosaurs: Simple answer tion of identity and homeotic transformation to a complex problem. Auk 119:1187–1201. by modulated BMP signaling. Science 289:438–441. Feduccia, A. 2006. Mesozoic aviary takes form. Pro­ de Beer, G. R. 1937. The Development of the Vertebrate ceedings of the National Academy of Sciences USA Skull. Clarendon Press, Oxford, United Kingdom. 103:5–6. de Beer, G. R., and E. J. W. Barrington. 1934. The seg­ Feduccia, A., T. Lingham-Soliar, and J. R. Hinch- mentation and chondrification of the skull of the liffe. 2005. Do feathered dinosaurs exist? Testing . Philosophical Transactions of the Royal Soci­ the hypothesis on neontological and paleontologi­ ety of London, Series B 223:411–467. cal evidence. Journal of Morphology 266:125–166. de Pinna, M. C. C. 1991. Concepts and tests of homology Feduccia, A., L. D. Martin, and S. Tarsitano. 2007. in the cladistic paradigm. Cladistics 7:367–394. Archaeopteryx 2007: Quo vadis? Auk 124:373–380. Dingus, L., and T. Rowe. 1998. The Mistaken Extinc­ Feduccia, A., and J. Nowicki. 2002. The hand of birds tion: Dinosaur Evolution and the Origin of Birds. revealed by early embryos. Naturwissen­ W. H. Freeman, New York. schaften 89:391–393. Drossopoulou, G., K. E. Lewis, J. J. Sanz-Ezquerro, N. Feduccia, A., and R. Wild. 1993. Birdlike characters in Nikbahkt, A. P. McMahon, C. Hofmann, and C. the Triassic archosaur Megalancosaurus. Naturwis­ Tickle. 2000. A model for anteroposterior pattern­ senschaften 80:564–566. ing of the vertebrate limb based on sequential long- Felsenstein, J. 2001. The troubled growth of statistical and short-range Shh signalling and Bmp signalling. . Systematic Biology 50:465–467. Development 127:1337–1348. Felsenstein, J. 2004. Inferring Phylogenies. Sinauer Elzanowski, A. 1999. A comparison of the jaw skeleton Associates, Sunderland, Massachusetts. in theropods and birds, with a description of the Forster, C. A., S. D. Sampson, L. M. Chiappe, and palate in the . Pages 311–323 in Avian D. W. Krause. 1998. The theropod ancestry of Paleontology at the Close of the 20th Century: Pro­ birds: New evidence from the Late Cretaceous of ceedings of the 4th International Meeting of the Soci­ Madagascar. Science 279:1915–1919. ety of Avian Paleontology and Evolution (S. L. Olson, Galis, F., M. Kundrát, and J. A. J. Metz. 2005. Hox genes, Ed.). Smithson Contributions in Paleobiology, no. 89. digit identities, and the theropod/bird transition. Elzanowski, A. 2002. (Upper Juras­ Journal of Experimental Zoology B 304:198–205. sic of ). Pages 129–159 in Mesozoic Birds: Galis, F., M. Kundrát, and B. Sinervo. 2003. An old Above the Heads of Dinosaurs (L. M. Chiappe and controversy solved: Bird embryos have five fingers. L. M. Witmer, Eds.). University of California Press, Trends in Ecology and Evolution 18:7–9. Berkeley. Galton, P. M. 1974. The ornithischian dinosaur Hyp- Elzanowski, A., and P. M. Galton. 1991. Braincase of silophodon from the Wealdon of the Isle of Wight. Enaliornis, an bird from England. Bulletin of the British Museum (Natural History), Journal of Vertebrate Paleontology 11:90–107. Series 251:1–152. Elzanowski, A., and P. Wellnhofer. 1996. Cranial Galton, P. M., and L. D. Martin. 2002. Enaliornis, an morphology of Archaeopteryx: Evidence from the Early Cretaceous hesperornithiform bird from Eng­ seventh skeleton. Journal of Vertebrate Paleontol­ land with comments on other Hesperornithiformes. ogy 16:81–94. Pages 317–338 in Mesozoic Birds: Above the Heads Ewer, R. F. 1965. The anatomy of the thecodont reptile of Dinosaurs (L. M. Chiappe and L. M. Witmer, Eds.). Euparkeria capensis Broom. Philosophical Trans­ University of California Press, Berkeley. actions of the Royal Society of London, Series B Gauthier, J. 1986. Saurischian monophyly and the ori­ 248:379–435. gin of birds. Memoirs of the California Academy of Feduccia, A. 1985. On why the dinosaur lacked Sciences, no. 8. feathers. Pages 75–79 in The Beginnings of Birds: Gauthier, J., and L. F. Gall, Eds. 2001. New Perspec­ Proceedings of the International Archaeopteryx Con­ tives on the Origin and Early Evolution of Birds: ference Eichstatt 1984 (M. K. Hecht, J. H. Ostrom, G. Proceedings of the International Symposium in Viohl, and P. Wellnhofer, Eds.). Freunde des Jura- Honor of John H. Ostrom. Yale University Press, Museums Eichstatt, Eichstatt. New Haven, Connecticut. Feduccia, A. 1993. Aerodynamic model for the early Gauthier, J. A., and K. Padian. 1985. Phylogenetic, evolution of feathers provided by Propithecus (Pri­ functional, and aerodynamic analyses of the ori­ mates, Lemuridae). Journal of Theoretical Biology gin of birds and their flight. Pages 185–197 in The 160:159–164. Beginnings of Birds: Proceedings of the International Feduccia, A. 1995. The aerodynamic model for the evo­ Archaeopteryx Conference Eichstatt 1984 (M. K. Hecht, lution of feathers and feather misinterpretation. J. H. Ostrom, G. Viohl, and P. Wellnhofer, Eds.). Courier Forschungsinstitut Senckenberg 181:65–77. Freunde des Jura-Museums Eichstatt, Eichstatt.

OM66_Text.indd 42 3/31/09 5:55:11 PM origin of birds 43

Gishlick, A. D. 2001. The function of the manus and Haughton, S. H. 1924. On a new type of thecodont from forelimb of Deinonychus antirrhopus and its impor­ the Middle Beaufort beds. Annals of the Transvaal tance for the origin of avian flight. Pages 301–318in Museum 11:93–97. New Perspectives on the Origin and Early Evolu­ Heckert, A. B., and S. G. Lucas. 2003. Clarifying aeto­ tion of Birds: Proceedings of the International Sym­ saur phylogeny requires more fossils, not more posium in Honor of John H. Ostrom (J. Gauthier trees—Reply to Intraorganismal homology, char­ and L. F. Gall, Eds.). Yale University Press, New acter construction, and the phylogeny of the aeto­ Haven, Connecticut. saurian archosaurs (Reptilia: Diapsida). Systematic Göhlich, U. B., and L. M. Chiappe. 2006. A new car­ Biology 52:253–255. nivorous dinosaur from the Solnhofen Heilmann, G. 1926. The Origin of Birds. Witherby, Lon­ archipelago. Nature 440:329–332. don. Goldman, N., J. P. Anderson, and A. G. Rodrigo. Hempel, C. G. 1966. Philosophy of Natural Science. 2000. Likelihood-based tests of topologies in phylo­ Prentice-Hall, Englewood Cliffs, New Jersey. genetics. Systematic Biology 49:652–670. Hinchliffe, J. R. 1985. “One, two, three” or “Two, three, Goloboff, P. A. 1999. NONA, version 1.9. Published four”: An embryologist’s view of the homologies by the author, Fundación e Instituto Miguel Lillo, of the digits and carpus of modern birds. Pages Tucuman, Argentina. [Online.] Available at www. 141–147 in The Beginnings of Birds: Proceedings of cladistics.com/aboutNona.htm. the International Archaeopteryx Conference Eichst­ Gosliner, T. M., and M. T. Ghiselin. 1984. Paral­ att 1984 (M. K. Hecht, J. H. Ostrom, G. Viohl, and lel evolution in opisthobranch gastropods and its P. Wellnhofer, Eds.). Freunde des Jura-Museums implications for phylogenetic methodology. Sys­ Eichstatt, Eichstatt. tematic Zoology 33:255–274. Hinchliffe, J. R., and M. K. Hecht. 1984. Homology Gould, S. J., and E. S. Vrba. 1982. Exaptation—A missing of the bird wing skeleton: Embryological versus term in the science of form. Paleobiology 8:4–15. paleontological evidence. Gower, D. J. 1997. The braincase of the early archosau­ 18:21–39. rian reptile Erythrosuchus africanus. Journal of Zool­ Holder, M. T., J. Sukumaran, and P. O. Lewis. 2008. ogy (London) 242:557–576. A justification for reporting the majority-rule con­ Gower, D. J. 2001. Possible postcranial pneumaticity in sensus tree in Bayesian phylogenetics. Systematic the last common ancestor of birds and crocodilians: Biology 57:814–821. Evidence from Erythrosuchus and other Mesozoic Holmgren, N. 1955. Studies on the phylogeny of birds. archosaurs. Naturwissenschaften 88:119–122. Acta Zoologica 36:243–328. Gower, D. J. 2002. Braincase evolution in suchian Holtz, T. R., Jr. 2000. A new phylogeny of the carnivo­ archosaurs (Reptilia: Diapsida): Evidence from rous dinosaurs. Gaia 15:5–61. the rauisuchian kupferzellensis. Holtz, T. R., Jr. 2001. Arctometatarsalia revisited: the Zoological Journal of the Linnean Society 136:49– problem of homoplasy in reconstructing theropod 76. phylogeny. Pages 99–122 in New Perspectives on Gower, D. J. 2003. Osteology of the early archosaurian the Origin and Early Evolution of Birds: Proceed­ reptile Erythrosuchus africanus Broom. Annals of the ings of the International Symposium in Honor of South African Museum 110:1–88. John H. Ostrom (J. Gauthier and L. F. Gall, Eds.). Gower, D. J., and A. D. Walker. 2002. New data on the Yale University Press, New Haven, Connecticut. braincase of the aetosaurian archosaur (Reptilia: Holtz, T. R., Jr. 2004. Tyrannosauroidea. Pages 111–136 Diapsida) Stagonolepis robertsoni Agassiz. Zoologi­ in The Dinosauria, 2nd ed. (D. B. Weishampel, P. cal Journal of the Linnean Society 136:7–23. Dodson, and H. Osmólska, Eds.). University of Gower, D. J., and E. Weber. 1998. The braincase of Eupark- California Press, Berkeley. eria, and the evolutionary relationships of birds and Holtz, T. R., Jr., R. E. Molnar, and P. J. Currie. 2004. crocodilians. Biological Reviews 73:367–411. Basal . Pages 71–110 in The Dinosauria, 2nd Gower, D. J., and M. Wilkinson. 1996. Is there any ed. (D. B. Weishampel, P. Dodson, and H. Osmólska, consensus on basal archosaur phylogeny? Pro­ Eds.). University of California Press, Berkeley. ceedings of the Royal Society of London, Series B Hou, L. H., L. D. Martin, Z. Zhou, and A. Feduccia. 263:1399–1406. 1999a. Archaeopteryx to opposite birds—Missing Hall, B. K. 2001. Development of the clavicles in birds link from the Mesozoic of China. Vertebrata PalAsi­ and . Journal of Experimental Zoology atica 37:88–95. 289:153–161. Hou, L., L. D. Martin, Z. Zhou, A. Feduccia, and F. Harris, J. D., and A. Downs. 2002. A drepanosaurid Zhang. 1999b. A diapsid skull in a new of the pectoral girdle from the (Whitaker) primitive bird Confuciusornis. Nature 399:679–682. Coelophysis quarry (Chinle Group, Rock Point For­ Houck, M. A., J. A. Gauthier, and R. E. Strauss. 1990. mation, Rhaetian), New Mexico. Journal of Verte­ Allometric scaling in the earliest fossil bird, Archae- brate Paleontology 22:70–75. opteryx lithographica. Science 247:195–198.

OM66_Text.indd 43 3/31/09 5:55:11 PM 44 ORNITHOLOGICAL MONOGRAPHS NO. 66

Hurum, J. H. 2001. Lower jaw of Gallimimus bullatus. Kirkland, J. I., L. E. Zanno, S. D. Sampson, J. M. Pages 34–41 in Mesozoic Vertebrate Life: New Clark, and D. D. DeBlieux. 2005b. A primitive Research Inspired by the Paleontology of Philip J. therizinosauroid dinosaur from the Early Creta­ Currie (D. H. Tanke and K. Carpenter, Eds.). Indi­ ceous of Utah. Nature 435:84–87. ana University Press, Bloomington. Kishino, H., and M. Hasegawa. 1989. Evaluation of Hwang, S., M. A. Norell, Q. Ji, and K. Gao. 2002. New the maximum likelihood estimate of the evolution­ specimens of Microraptor zhaoianus (Theropoda: ary tree topologies from DNA sequence data, and Dromaeosauridae) from northeastern China. Amer­ the branching order in Hominoidea. Journal of ican Museum Novitates 3381:1–44. 29:170–179. Hwang, S., M. A. Norell, Q. Ji, and K. Gao. 2004. A Kluge, A. G. 1997. Testability and the refutation and large compsognathid from the Early Cretaceous corroboration of cladistic hypotheses. Cladistics of China. Journal of Systematic 13:81–96. Palaeontology 2:13–30. Kluge, A. G. 1999. The science of phylogenetic system­ Iordansky, N. N. 1973. The skull of the . Pages atics: Explanation, prediction, and test. Cladistics 201–262 in Biology of the Reptilia, vol. 4 (C. Gans 15:429–436. and T. S. Parsons, Eds.). Academic Press, London. Kluge, A. G. 2001. Philosophical conjectures and their Jenner, R. A. 2004. The scientific status of metazoan cla­ refutation. Systematic Biology 50:322–330. distics: Why current research practice must change. Kobayashi, Y., and R. Barsbold. 2005. Anatomy of Zoologica Scripta 33:293–310. Harpymimus okladnikovi Barsbold and Perle 1984 Ji, Q., P. J. Currie, M. A. Norell, and S. Ji. 1998. Two (Dinosauria: Theropoda) of Mongolia. Pages 97– feathered dinosaurs from northeastern China. 126 in The Carnivorous Dinosaurs (K. Carpenter, Nature 393:753–761. Ed.). Indiana University Press, Bloomington. Ji, Q., M. A. Norell, K. Q. Gao, S. Ji, and D. Ren. 2001. Kobayashi, Y., and J. Lü. 2003. A new ornithomimid The distribution of integumentary structures in a dinosaur with gregarious habits from the Late Cre­ . Nature 410:1084–1088. taceous of China. Acta Palaeontologica Polonica Ji, Q., and C. Yuan. 2002. Discovery of two kinds of pro­ 48:235–259. tofeathered pterosaurs in the Mesozoic Daohugou Kuhn, T. S. 1970. The Structure of Scientific Revolu­ biota in the Ningcheng region and its stratigraphic tions, 2nd ed. University of Chicago Press, Chicago, and biologic significances. [In Chinese.] Geological Illinois. Review 48:221–224. Kundrát, M. 2008. Primary chondrification foci in the Jones, T. D., J. A. Ruben, P. F. A. Maderson, and L. D. wing basipodium of Struthio camelus with com­ Martin. 2001. Longisquama fossil and feather mor­ ments on interpretation of autopodial elements in phology: Response. Science 291:1901–1902. Crocodilia and Aves. Journal of Experimental Zool­ Jones, T. D., J. A. Ruben, L. D. Martin, E. N. Kuro- ogy B 310:1–12. chkin, A. Feduccia, P. F. A. Maderson, W. J. Kundrát, M., V. Seichert, A. P. Russell, and K. Hillenius, N. R. Geist, and V. Alifanov. 2000. Smetana, Jr. 2002. Pentadactyl pattern of the avian Non-avian feathers in a archosaur. Sci­ wing autopodium and pyramid reduction hypothe­ ence 288:2202–2205. sis. Journal of Experimental Zoology B 294:152–159. Kearney, M. 2002. Fragmentary taxa, missing data, and Kurochkin, E. N. 2006a. Basal diversification of Aves. ambiguity: Mistaken assumptions and conclusions. Pages 219–232 in Evolution of the Biosphere and Systematic Biology 51:369–381. Biodiversity: To 70 Years of A. Yu. Rozanov (S. V. Kearney, M., and J. M. Clark. 2003. Problems due to Rozhnov, Ed.). [In Russian.] Toravirtzestvo Scien­ missing data in phylogenetic analyses including tific Editions KMK, Moscow. fossils: A critical review. Journal of Vertebrate Pale­ Kurochkin, E. N. 2006b. Parallel evolution of theropod ontology 23:263–274. dinosaurs and birds. [In Russian.] Zoologicheskii Kearney, M., and O. Rieppel. 2006. Rejecting “the Zhurnal 85:283–297 [English version, Entomologi­ given” in systematics. Cladistics 22:369–377. cal Review 86, S45–S58]. Kirkland, J. I., D. K. Smith, and D. G. Wolfe. 2005a. Kurzanov, S. M. 1982. Structural characteristics of the braincase of Nothronychus mckinleyi Kirk­ forelimbs of Avimimus. Paleontological Journal land and Wolfe 2001 (Theropoda; Therizinosau­ 16:108–112. roidea) from the Upper Cretaceous (Turonian) of Kurzanov, S. M. 1983. New data on the pelvic structure west-central New Mexico. Pages 87–96 in The Car­ of Avimimus. Paleontological Journal 17:110–111. nivorous Dinosaurs (K. Carpenter, Ed.). Indiana Kurzanov, S. M. 1985. The skull structure of the dino­ University Press, Bloomington. saur Avimimus. Paleontological Journal 19:92–99. Kirkland, J. I., and D. G. Wolfe. 2001. First defini­ Langer, M. C. 2004. Basal Saurischia. Pages 25–46 in tive therizinosaurid (Dinosauria; Theropoda) from The Dinosauria, 2nd ed. (D. B. Weishampel, P. Dod­ North America. Journal of Vertebrate Paleontology son, and H. Osmólska, Eds.). University of Califor­ 21:410–414. nia Press, Berkeley.

OM66_Text.indd 44 3/31/09 5:55:12 PM origin of birds 45

Larsson, H. C. E., and G. P. Wagner. 2002. Pentadactyl Madsen, J. H., Jr. 1993. Allosaurus fragilis: A revised ground state of the avian wing. Journal of Experi­ osteology. [Reprint.] Utah Geological Survey Bul­ mental Zoology B 294:146–151. letin, no. 109. Larssson, H. C. E., and G. P. Wagner. 2003. Old mor­ Madsen, J. H., Jr., and S. P. Welles. 2000. Ceratosau- phologies misinterpreted. Trends in Ecology and rus (Dinosauria, Theropoda): A Revised Osteology. Evolution 18:10. Utah Geological Survey, Salt Lake City. Lingham-Soliar, T. 2003a. Evolution of birds: Ichthyo­ Makovicky, P., and P. J. Currie. 1998. The presence of saur integumental fibers conform to dromaeosaur a furcula in tyrannosaurid theropods, and its phy­ protofeathers. Naturwissenschaften 90:428–432. logenetic and functional implications. Journal of Lingham-Soliar, T. 2003b. The dinosaurian origin of Vertebrate Paleontology 18:143–149. feathers: Perspectives from dolphin (Cetacea) col­ Makovicky, P. J., and G. J. Dyke. 2001. Naive falsification lagen fibers. Naturwissenschaften 90:563–567. and the origin of birds. Pages 501–510 in New Perspec­ Lingham-Soliar, T., A. Feduccia, and X. Wang. 2007. tives on the Origin and Early Evolution of Birds: Pro­ A new Chinese specimen indicates that ‘protofeath­ ceedings of the International Symposium in Honor of ers’ in the Early Cretaceous theropod dinosaur John H. Ostrom (J. Gauthier and L. F. Gall, Eds.). Yale Sinosauropteryx are degraded collagen fibres. Pro­ University Press, New Haven, Connecticut. ceedings of the Royal Society of London, Series B Makovicky, P., Y. Kobayashi, and P. J. Currie. 2004. 274:1823–1829. Ornithomimosauria. Pages 137–150 in The Dino­ Lingham-Soliar, T., and J. Wesley-Smith. 2008. First sauria, 2nd ed. (D. B. Weishampel, P. Dodson, and investigation of the collagen D-band ultrastructure H. Osmólska, Eds.). University of California Press, in fossilized vertebrate integument. Proceedings Berkeley. of the Royal Society of London, Series B 275:2207– Makovicky, P., and M. A. Norell. 2004. Troodontidae. 2212. Pages 184–195 in The Dinosauria, 2nd ed. (D. B. Livezey, B. C. 1998. A phylogenetic analysis of the Grui­ Weishampel, P. Dodson, and H. Osmólska, Eds.). formes (Aves) based on morphological characters, University of California Press, Berkeley. with an emphasis on the rails (Rallidae). Philosoph­ Makovicky, P. J., M. A. Norell, J. M. Clark, and T. ical Transactions of the Royal Society of London, Rowe. 2003. Osteology and relationships of Byrono- Series B 353:2077–2151. saurus jaffei (Theropoda: Troodontidae). American Livezey, B. C. 2003. Evolution of flightlessness in rails Museum Novitates 3402:1–32. (: Rallidae): Phylogenetic, ecomorpho­ Makovicky, P., and H.-D. Sues. 1998. Anatomy and logical, and ontogenetic perspectives. Ornithologi­ phylogenetic relationships of the theropod dinosaur cal Monographs, no. 53. Microvenator celer from the Lower Cretaceous of Livezey, B. C., and R. L. Zusi. 2006. Higher-order phy­ Montana. American Museum Novitates 3240:1–27. logeny of modern birds (Theropoda, Aves: Neorni­ Marsh, O. C. 1880. Odontornithes: A Monograph on thes) based on : I. Methods the Extinct Toothed Birds of North America. United and characters. Bulletin of the Carnegie Museum of States Geological Exploration of the 40th Parallel. Natural History, no. 37. U.S. Government Printing Office,W ashington, D.C. Longrich, N. 2006. Structure and function of hindlimb Martin, L. D. 1983. The origin of birds and of avian feathers in Archaeopteryx lithographica. Paleobiology flight. Pages 105–129 in Current , vol. 1 32:417–431. (R. F. Johnston, Ed.). Plenum Press, New York. Lü, J. 2002. A new oviraptorosaurid (Theropoda: Ovi­ Martin, L. D. 1985. The relationship of Archaeopteryx raptorosauria) from the Late Cretaceous of south­ to other birds. Pages 177–183 in The Beginnings of ern China. Journal of Vertebrate Paleontology Birds: Proceedings of the International Archaeop- 22:871–875. teryx Conference Eichstatt 1984 (M. K. Hecht, J. H. Lü, J., Z. Dong, Y. Azuma, R. Barsbold, and Y. Tomida. Ostrom, G. Viohl, and P. Wellnhofer, Eds.). Freunde 2002. Oviraptorosaurs compared to birds. Pages des Jura-Museums Eichstatt, Eichstatt. 175–189 in Proceedings of the 5th Symposium of Martin, L. D. 1991. Mesozoic birds and the origin the Society of Avian Paleontology and Evolution (Z. of birds. Pages 485–540 in Origins of the Higher Zhou and F. Zhang, Eds.). Science Press, . Groups of : Controversy and Consensus Lü, J., D. Huang, and L. Qiu. 2005. The pectoral girdle (H.-P. Schultze and L. Trueb, Eds.). Cornell Univer­ and the forelimb of Heyuannia (Dinosauria: Ovi­ sity Press, Ithaca, New York. raptorosauria). Pages 256–273 in The Carnivorous Martin, L. D. 1995. A new skeletal model of Archaeop- Dinosaurs (K. Carpenter, Ed.). Indiana University teryx. Archaeopteryx 13:33–40. Press, Bloomington. Martin, L. D. 2004. A basal archosaurian origin of Maderson, P. F. A., and L. Alibardi. 2000. The devel­ birds. Acta Zoologica Sinica 50:978–990. opment of the sauropsid integument: A contribu­ Martin, L. D., and O. Bonner. 1977. An immature tion to the problem of the origin and evolution of specimen of Baptornis advenus from the Cretaceous feathers. American Zoologist 40:513–529. of Kansas. Auk 94:787–789.

OM66_Text.indd 45 3/31/09 5:55:12 PM 46 ORNITHOLOGICAL MONOGRAPHS NO. 66

Martin, L. D., and S. A. Czerkas. 2000. The fossil Naylor, G. J. P., and D. C. Adams. 2001. Are the fossil record of feather evolution in the Mesozoic. Ameri­ data really at odds with the molecular data? Mor­ can Zoologist 40:687–694. phological evidence for Cetartiodactyla phylogeny Martin, L. D., V. L. Naples, and Z. Zhou. 1998a. The reexamined. Systematic Biology 50:444–453. furcula in early birds. Journal of Vertebrate Paleon­ Nesbitt, S. J. 2007. The anatomy of Effigia okeeffeae tology 18(Supplement 3):60A–61A. (Archosauria, Suchia), theropod-like convergence, Martin, L. D., and J. D. Stewart. 1977. Teeth in Ichthy- and the distribution of related taxa. Bulletin of the ornis (: Aves). Nature 195:1331–1332. American Museum of Natural History 302:1–84. Martin, L. D., and J. D. Stewart. 1985. Homologies in Nesbitt, S. J., and M. A. Norell. 2006. Extreme conver­ the avian tarsus. Nature 315:159–160. gence in the body plans of an early suchian (Archo­ Martin, L. D., and J. D. Stewart. 1999. Implantation sauria) and ornithomimid dinosaurs (Theropoda). and replacement of bird teeth. Smithsonian Contri­ Proceedings of the Royal Society of London, Series butions to Paleobiology 89:295–300. B 273:1045–1048. Martin, L. D., J. D. Stewart, and K. N. Whetstone. Nicholls, E. L., and A. P. Russell. 1985. Structure 1980. The origin of birds: Structure of the tarsus and and function of the pectoral girdle and forelimb of teeth. Auk 97:86–93. altus (Theropoda: ). Martin, L. D., and J. Tate, Jr. 1976. The skeleton of Palaeontology 28:643–677. Baptornis advenus. Smithsonian Contributions to Nixon, K. C., and J. M. Carpenter. 1993. On out­ Paleobiology 27:35–66. groups. Cladistics 9:413–426. Martin, L. D., and Z. Zhou. 1997. Archaeopteryx-like Norell, M. A., J. Clark, and P. J. Makovicky. 2001. skull in enantiornithine bird. Nature 389:556. Phylogenetic relationships among coelurosaurian Martin, L. D., Z. Zhou, L. Hou, and A. Feduccia. theropods. Pages 49–67 in New Perspectives on the 1998b. Confuciusornis sanctus compared to Archaeop- Origin and Early Evolution of Birds: Proceedings teryx lithographica. Naturwissenschaften 85:286–289. of the International Symposium in Honor of John Marya ´nska, T., H. Osmólska, and M. Wolsan. 2002. H. Ostrom (J. Gauthier and L. F. Gall, Eds.). Yale Avialan status for Oviraptorosauria. Acta Palaeon­ University Press, New Haven, Connecticut. tologica Polonica 47:97–116. Norell, M. A., and M. Ellison. 2005. Unearthing the Mayr, G., B. Pohl, S. Hartman, and D. S. Peters. 2007. Dragon: The Great Feathered Dinosaur Discovery. The tenth skeletal specimen of Archaeopteryx. Zoo­ Pi Press, New York. logical Journal of the Linnean Society 149:97–116. Norell, M. A., and S. H. Hwang. 2004. A troodon­ Mayr, G., B. Pohl, and D. S. Peters. 2005. A well-pre­ tid dinosaur from Ukhaa Tolgod (Late Creta­ served Archaeopteryx specimen with theropod fea­ ceous Mongolia). American Museum Novitates tures. Science 310:1483–1486. 3446:1–9. McDowell, S. 1978. Homology mapping of the primi­ Norell, M. A., and P. J. Makovicky. 1997. Important tive archosaurian reptile palate on the palate of features of the dromaeosaur skeleton: Information birds. Evolutionary Theory 4:81–94. from a new specimen. American Museum Novi­ McGowan, C. 1984. Evolutionary relationships of tates 3215:1–28. and carinates: Evidence from ontogeny of Norell, M. A., and P. J. Makovicky. 1999. Important the tarsus. Nature 307:733–735. features of the dromaeosaurid skeleton II: Informa­ McGowan, C. 1985. Homologies in the avian tarsus. tion from newly collected specimens of Velociraptor Nature 315:160. mongoliensis. American Museum Novitates 3282:1– Montagna, W. 1945. A re-investigation of the develop­ 43. ment of the wing of the . Journal of Morphol­ Norell, M. A., and P. J. Makovicky. 2004. Dromaeo­ ogy 76:87–113. sauridae. Pages 196–209 in The Dinosauria, 2nd ed. Mook, C. C. 1921. Notes on the postcranial skeleton in (D. B. Weishampel, P. Dodson, and H. Osmólska, the Crocodilia. Bulletin of the American Museum Eds.). University of California Press, Berkeley. of Natural History 44:67–99. Norell, M. A., P. J. Makovicky, and J. M. Clark. 2000. Morse, E. S. 1872. On the tarsus and carpus of birds. A new troodontid from Ukhaa Tolgod, Mongolia. Annals of the Lyceum of Natural History 10:1–22. Journal of Vertebrate Paleontology 20:7–11. Murray, P. F., and P. Vickers-Rich. 2004. Magnificent Norell, M. A., and X. Xu. 2005. Feathered dinosaurs. Mihirungs: The Colossal Flightless Birds of the Annual Review of Earth and Planetary Sciences Australian Dreamtime. Indiana University Press, 33:277–299. Bloomington. Norman, D. B. 1990. Problematic Theropoda: “Coelu­ Naples, V. L., L. D. Martin, and J. Simmons. 2002. The rosaurs.” Pages 280–305 in The Dinosauria (D. B. pelvis in early birds and dinosaurs. Pages 203–210 Weishampel, P. Dodson, and H. Osmólska, Eds.). in Proceedings of the 5th Symposium of the Society University of California Press, Berkeley. of Avian Paleontology and Evolution (Z. Zhou and Novas, F. E. 1993. New information on the system­ F. Zhang, Eds.). Science Press, Beijing. atics and postcranial skeleton of Herrerasaurus

OM66_Text.indd 46 3/31/09 5:55:13 PM origin of birds 47

ischigualastensis (Theropoda: ) Padian, K. 2001a. Stages in the origin of bird flight: from the (Upper Triassic) Beyond the arboreal–cursorial dichotomy. Pages of Argentina. Journal of Vertebrate Paleontology 255–272 in New Perspectives on the Origin and 13:400–423. Early Evolution of Birds: Proceedings of the Inter­ Novas, F. E., and D. Pol. 2002. Alvarezsaurid rela­ national Symposium in Honor of John H. Ostrom tionships reconsidered. Pages 121–125 in Meso­ (J. Gauthier and L. F. Gall, Eds.). Yale University zoic Birds: Above the Heads of Dinosaurs (L. M. Press, New Haven, Connecticut. Chiappe and L. M. Witmer, Eds.). University of Padian, K. 2001b. The false issues of bird origins: An California Press, Berkeley. historiographic perspective. Pages 485–499 in New Novas, F. E., and P. F. Puerta. 1997. New evidence con­ Perspectives on the Origin and Early Evolution cerning avian origins from the Late Cretaceous of of Birds: Proceedings of the International Sympo­ Patagonia. Nature 387:390–392. sium in Honor of John H. Ostrom (J. Gauthier and Osmólska, H., P. J. Currie, and R. Barsbold. 2004. L. F. Gall, Eds.). Yale University Press, New Haven, Oviraptorosauria. Pages 165–183 in The Dinosau­ Connecticut. ria, 2nd ed. (D. B. Weishampel, P. Dodson, and H. Padian, K. 2004. Basal Avialae. Pages 210–231 in The Osmólska, Eds.). University of California Press, Dinosauria, 2nd ed. (D. B. Weishampel, P. Dodson, Berkeley. and H. Osmólska, Eds.). University of California Osmólska, H., E. Roniewicz, and R. Barsbold. 1972. Press, Berkeley. A new dinosaur, Gallimimus bullatus n. gen., n. sp. Padian, K., and L. M. Chiappe. 1998. The origin (Ornithomimidae) from the Upper Cretaceous of and early evolution of birds. Biological Reviews Mongolia. Palaeontologia Polonica 27:103–143. 73:1–42. Ostrom, J. H. 1969. Osteology of Deinonychus antirrho- Padian, K., Q. Ji, and S. Ji. 2001. Feathered dinosaurs and pus, an unusual theropod from the Lower Creta­ the origin of flight. Pages 117–135in Mesozoic Verte­ ceous of Montana. Bulletin of the Peabody Museum brate Life: New Research Inspired by the Paleontol­ of Natural History, no. 30. ogy of Philip J. Currie (D. H. Tanke and K. Carpenter, Ostrom, J. H. 1973. The ancestry of birds. Nature Eds.). Indiana University Press, Bloomington. 242:136. Parker, W. K. 1891. On the morphology of a reptilian Ostrom, J. H. 1974. The pectoral girdle and forelimb bird Opisthocomus cristatus. Transactions of the Zoo­ function of Deinonychus (Reptilia: Saurischia): A logical Society of London 13:43–85. correction. Postilla 165:1–11. Parrish, J. M. 1993. Phylogeny of the Crocodylotarsi, Ostrom, J. H. 1975. The origin of birds. Annual Review with reference to archosaurian and crurotarsan of Earth and Planetary Sciences 3:55–77. monophyly. Journal of Vertebrate Paleontology Ostrom, J. H. 1976a. Archaeopteryx and the origin of 13:287–308. birds. Biological Journal of the Linnean Society Paul, G. S. 2002. Dinosaurs of the Air: The Evolution 8:91–182. and Loss of Flight in Dinosaurs and Birds. Johns Ostrom, J. H. 1976b. Some hypothetical anatomical Hopkins University Press, Baltimore, Maryland. stages in the evolution of avian flight. Smithsonian Pérez-Moreno, B. P., J. L. Sanz, A. D. Buscalioni, J. J. Contributions to Paleobiology 27:1–21. Moratalla, F. Ortega, and D. Rasskin-Gutman. Ostrom, J. H. 1978. The osteology of Compsognathus lon- 1994. A unique multitoothed ornithomimosaur gipes Wagner. Zitteliana 4:73–118. dinosaur from the Lower Cretaceous of Spain. Ostrom, J. H. 1985. The meaning of Archaeopteryx. Nature 370:363–367. Pages 161–176 in The Beginnings of Birds: Proceed­ Perle, A., L. M. Chiappe, R. Barsbold, J. M. Clark, ings of the International Archaeopteryx Conference and M. A. Norell. 1994. Skeletal morphology of Eichstatt 1984 (M. K. Hecht, J. H. Ostrom, G. Viohl, Mononykus olecranus (Theropoda: Avialae) from the and P. Wellnhofer, Eds.). Freunde des Jura Muse­ Late Cretaceous of Mongolia. American Museum ums Eichstatt, Eichstatt. Novitates 3105:1–29. Owen, R. 1850. On the communication between the Perle, A., M. A. Norell, L. M. Chiappe, and J. M. cavity of the tympanum and the palate in the Clark 1993. Flightless bird from the Cretaceous of Crocodilia (gavials, alligators and crocodiles). Mongolia. Nature 362:623–626. Philosophical Transactions of the Royal Society of Peyer, K. 2006. A reconsideration of Compsognathus London 140:521–527. from the Upper Tithonian of Canjuers, southeast­ Padian, K. 1982. and the origin of ern France. Journal of Vertebrate Paleontology major adaptations: Vertebrate flight as a paradigm 24:879–896. for the analysis of patterns. Pages 387–392 in Pro­ Popper, K. R. 1998. Conjectures and Refutations: The ceedings of the Third North American Paleonto­ Growth of Scientific Knowledge, 5th ed.R outledge, logical Convention, vol. 2 (B. L. Mamet and M. J. New York. Copeland, Eds.). Business and Economic Service, Popper, K. R. 2002. The Logic of Scientific Discovery. Toronto, Ontario. Routledge, New York.

OM66_Text.indd 47 3/31/09 5:55:13 PM 48 ORNITHOLOGICAL MONOGRAPHS NO. 66

Prum, R. O. 1999. Development and evolutionary ori­ Sanderson, M. J., and H. B. Shaffer. 2002. Trouble­ gin of feathers. Journal of Experimental Zoology B shooting molecular phylogenetic analyses. Annual 285:291–306. Review of Ecology and Systematics 33:49–72. Prum, R. O. 2002. Why ornithologists should care about Sanz, J. L., and J. F. Bonaparte. 1992. A new order of the theropod origin of birds. Auk 119:1–17. birds (Class Aves) from the Lower Cretaceous of Prum, R. O. 2003. Are current critiques of the theropod Spain. Natural History Museum of Los Angeles origin of birds science? Rebuttal to Feduccia (2002). County Scientific Series 36:39–49. Auk 120:550–561. Sanz, J. L., L. M. Chiappe, B. P. Pérez-Moreno, J. J. Prum, R. O., and A. H. Brush. 2002. The evolutionary Moratalla, F. Hernández-Carrasquilla, A. D. origin and diversification of feathers. Quarterly Buscalioni, F. Ortega, F. J. Poyato-Ariza, D. Review of Biology 77:261–295. Rasskin-Gutman, and X. Martínez-Delclòs. Prum, R. O., and J. Dyck. 2003. A hierarchical model 1997. A nestling bird from the Early Cretaceous of of : Morphology, development, and Spain: Implications for avian skull and neck evolu­ evolution. Journal of Experimental Zoology B tion. Science 276:1543–1546. 298:73–90. Sanz, J. L., B. P. Pérez-Moreno, L. M. Chiappe, and Prum, R. O. 2001. Longisquama fossil and feather mor­ A. D. Buscalioni. 2002. The birds from the Lower phology. Science 291:1899–1900. Cretaceous of Las Hoyas (Province of Cuenca, Spain). Raath, M. A. 1985. The theropod Syntarsus and its bear­ Pages 209–229 in Mesozoic Birds: Above the Heads ings on the origin of birds. Pages 219–227 in The of Dinosaurs (L. M. Chiappe and L. M. Witmer, Eds.). Beginnings of Birds: Proceedings of the International University of California Press, Berkeley. Archaeopteryx Conference Eichstatt 1984 (M. K. Hecht, Sawyer, R. H., and L. W. Knapp. 2003. Avian skin J. H. Ostrom, G. Viohl, and P. Wellnhofer, Eds.). Fre­ development and the evolutionary origin of feath­ unde des Jura-Museums Eichstatt, Eichstatt. ers. Journal of Experimental Zoology B 298:57–72. Reese, A. M. 1915. The Alligator and Its Allies. Knicker­ Sawyer, R. H., B. A. Salvatore, T. F. Potylicki, J. O. bocker Press, New York. French, T. C. Glenn, and L. W. Knapp. 2003a. Rieppel, O. 1985. The recessus scalae tympani and its Origin of feathers: Feather beta (β) keratins are bearing on the classification of reptiles. Journal of expressed in discrete epidermal populations of Herpetology 19:373–384. embryonic scutate scales. Journal of Experimental Rieppel, O., and M. Kearney. 2002. Similarity. Biologi­ Zoology B 295:12–24. cal Journal of the Linnean Society 75:59–82. Sawyer, R. H., L. D. Washington, B. A. Salvatore, T. C. Romanoff, A. L. 1960. The Avian Embryo: Structural Glenn, and L. W. Knapp. 2003b. Origin of archo­ and Functional Development. Macmillan, New saurian integumentary appendages: The bristles of York. the wild turkey beard express feather-type β kera­ Romer, A. 1956. Osteology of the Reptiles. University of tins. Journal of Experimental Zoology B 297:27–34. Chicago Press, Chicago, Illinois. Schweitzer, M. H. 2001. Evolutionary implications of Rose, M. R., and G. V. Lauder. 1996. Post- possible protofeather structures associated with a adaptationism. Pages 1–8 in Adaptation (M. R. specimen of Shuvuuia deserti. Pages 181–192 in New Rose and G. V. Lauder, Eds.). Academic Press, San Perspectives on the Origin and Early Evolution Diego, California. of Birds: Proceedings of the International Sympo­ Rowe, T., and J. Gauthier. 1990. Ceratosauria. Pages sium in Honor of John H. Ostrom (J. Gauthier and 151–168 in The Dinosauria (D. B. Weishampel, P. L. F. Gall, Eds.). Yale University Press, New Haven, Dodson, and H. Osmólska, Eds.). University of Connecticut. California Press, Berkeley. Schweitzer, M. H., J. A. Watt, R. Avci, L. Knapp, L. M. Ruben, J. A., and T. D. Jones. 2000. Selective factors Chiappe, M. A. Norell, and M. Marshall. 1999. associated with the origin of fur and feathers. Beta-keratin specific immunological reactivity in American Zoologist 40:585–596. feather-like structures of the Cretaceous alvarez­ Russell, A. P., and D. J. Joffe. 1985. The early develop­ saurid, Shuvuuia deserti. Journal of Experimental ment of the quail (Coturnix c. japonica) furcula recon­ Zoology B 285:146–157. sidered. Journal of Zoology (London) 206:69–81. Senter, P. 2004. Phylogeny of Drepanosauridae (Rep­ Russell, D. A., and Z. Dong. 1993a. A nearly complete tilia: Diapsida). Journal of Systematic Paleontology skeleton of a new troodontid dinosaur from the 2:257–268. Early Cretaceous of the Ordos Basin, Inner Mongo­ Senter, P., R. Barsbold, B. B. Britt, and D. A. Burn- lia, People’s Republic of China. Canadian Journal ham. 2004. Systematics and evolution of Dromaeo­ of Earth Sciences 30:2163–2173. sauridae (Dinosauria: Theropoda). Bulletin of the Russell, D. A., and Z. Dong. 1993b. The affinities of a Gunma Museum of Natural History 8:1–20. new theropod from the Alxa Desert, Inner Mongo­ Senter, P., and J. H. Robins. 2003. Taxonomic status lia, People’s Republic of China. Canadian Journal of the specimens of Archaeopteryx. Journal of Verte­ of Earth Sciences 30:2107–2127. brate Paleontology 23:961–965.

OM66_Text.indd 48 3/31/09 5:55:14 PM origin of birds 49

Sereno, P. C. 1991. Basal archosaurs: Phylogenetic rela­ C. Moritz, and B. K. Mable, Eds.). Sinauer Associ­ tionships and functional implications. Society for ates, Sunderland, Massachusetts. Vertebrate Paleontology Memoirs, no. 2. Tarsitano, S. 1991. Archaeopteryx: Quo vadis? Pages Sereno, P. C. 1993. The pectoral girdle and forelimb of 541–576 in Origins of the Higher Groups of Tet­ the basal theropod Herrerasaurus ischigualastensis. rapods: Controversy and Consensus (H.-P. Schul­ Journal of Vertebrate Paleontology 13:425–450. tze and L. Trueb, Eds.). Cornell University Press, Sereno, P. C. 1999. The evolution of dinosaurs. Science Ithaca, New York. 284:2137–2147. Tarsitano, S., and M. K. Hecht. 1980. A reconsideration Sereno, P. C. 2001. Alvarezsaurids: Birds or ornithomi­ of the reptilian relationships of Archaeopteryx. Zoo­ mosaurs? Pages 69–98 in New Perspectives on the logical Journal of the Linnean Society 69:149–182. Origin and Early Evolution of Birds: Proceedings Turner, A. H., D. Pol, J. A. Clarke, G. M. Erickson, of the International Symposium in Honor of John and M. A. Norell. 2007. A basal dromaeosaurid H. Ostrom (J. Gauthier and L. F. Gall, Eds.). Yale and size evolution preceding avian flight. Science University Press, New Haven, Connecticut. 317:1378–1381. Sereno, P. C. 2007. The phylogenetic relationships of Tykoski, R. S., and T. Rowe. 2004. Ceratosauria. Pages early dinosaurs: A comparative report. Histoire de 47–70 in The Dinosauria, 2nd ed. (D. B. Weisham­ la Biologie (Paris) 19:145–155. pel, P. Dodson, and H. Osmólska, Eds.). University Sereno, P. C., and A. B. Arcucci. 1994. Dinosaurian of California Press, Berkeley. precursors from the Middle Triassic of Argentina: Unwin, D. M. 2006. The Pterosaurs: From Deep Time. Marasuchus lilloensis, gen. nov. Journal of Vertebrate Pi Press, New York. Paleontology 14:53–73. Unwin, D. M., and N. N. Bakhurina. 1994. Sordes pilosus Sereno, P. C., and F. E. Novas. 1993. The skull and the nature of the flight apparatus. and neck of the basal theropod Herrerasaurus Nature 371:62–64. ischigualastensis. Journal of Vertebrate Paleontol­ Unwin, D. M., and M. J. Benton. 2001. Longisquama ogy 13:451–476. fossil and feather morphology. Science 291:1899. Sereno, P. C., C. Rao, and J. Li. 2002. Sinornis santensis Upchurch, P., P. M. Barrett, and P. Dodson. 2004. (Aves: Enantiornithes) from the early Cretaceous Sauropoda. Pages 259–324 in The Dinosauria, 2nd of northeastern China. Pages 184–208 in Meso­ ed. (D. B. Weishampel, P. Dodson, and H. Osmól­ zoic Birds: Above the Heads of Dinosaurs (L. M. ska, Eds.). University of California Press, Berkeley. Chiappe and L. M. Witmer, Eds.). University of Vargas, A. O., and J. F. Fallon. 2005a. Birds have dino­ California Press, Berkeley. saur : The molecular evidence. Journal of Sereno, P. C., J. A. Wilson, H. C. Larsson, D. B. Experimental Zoology B 304:85–89. Dutheil, and H.-D. Sues. 1994. Early Cretaceous Vargas, A. O., and J. F. Fallon. 2005b. The digits of the dinosaurs from the Sahara. Science 266:267–271. wing of birds are 1, 2, and 3. A review. Journal of Sharov, A. G. 1970. An unusual reptile from the Experimental Zoology B 304:206–219. Lower Triassic of Fergana. Paleontological Journal Vickers-Rich, P., L. M. Chiappe, and S. Kurzanov. 1970:112–116. 2002. The enigmatic birdlike dinosaur Avimimus Steel, M., and D. Penny. 2000. Parsimony, likelihood, portentosus. Pages 65–86 in Mesozoic Birds: Above and the role of models in molecular phylogenetics. the Heads of Dinosaurs (L. M. Chiappe and L. M. Molecular Biology and Evolution 17:839–850. Witmer, Eds.). University of California Press, Sues, H.-D., P. E. Olsen, J. G. Carter, and D. M. Scott. Berkeley. 2003. A new crocodylomorph archosaur from the Voigt, S., M. Buchwitz, J. Fischer, D. Krause, and R. Upper Triassic of North Carolina. Journal of Verte­ Georgi. 2009. Feather-like development of Trias­ brate Paleontology 23:329–343. sic diapsid skin appendages. Naturwissenschaften Suzuki, S., L. M. Chiappe, G. J. Dyke, M. Watabe, R. 96:81–86. Barsbold, and K. Tsogtbaatar. 2002. A new spec­ Wagner, G. P. 2005. The developmental evolution of imen of Shuvuuia deserti Chiappe et al. 1998 from avian digit homology: An update. Theory in Biosci­ the Mongolian Late Cretaceous with a discussion of ences 124:165–183. the relationships of alvarezsaurids to other thero­ Wagner, G. P., and J. A. Gauthier. 1999. 1,2,3 = 2,3,4: A pod dinosaurs. Natural History Museum of Los solution to the problem of the homology of the dig­ Angeles County Scientific Series, no. 494. its in the avian hand. Proceedings of the National Swofford, D. L. 2003. PAUP*: Phylogenetic Analysis Academy of Sciences USA 96:5111–5116. Using Parsimony (*and Other Methods), version Walker, A. D. 1961. Triassic reptiles from the Elgin 4.0b10. Sinauer Associates, Sunderland, Massa­ area: Stagonolepis, Dasygnathus and their allies. chusetts. Philosophical Transactions of the Royal Society of Swofford, D. L., G. J. Olsen, P. J. Waddell, and D. M. London, Series B 244:103–204. Hillis. 1996. Phylogenetic inference. Pages 407– Walker, A. D. 1964. Triassic reptiles from the Elgin 514 in Molecular Systematics, 2nd ed. (D. M. Hillis, area: Ornithosuchus and the origin of carnosaurs.

OM66_Text.indd 49 3/31/09 5:55:14 PM 50 ORNITHOLOGICAL MONOGRAPHS NO. 66

Philosophical Transactions of the Royal Society of Wiens, J. J., P. T. Chippendale, and D. M. Hillis. 2003. London, Series B 248:53–134. When are phylogenetic analyses misled by conver­ Walker, A. D. 1970. A revision of the Jurassic reptile gence? A case study in Texas cave salamanders. Hallopus victor (Marsh), with remarks on the classi­ Systematic Biology 52:501–514. fication of crocodiles. Philosophical Transactions of Wilkinson, M. 1996. Majority-rule reduced consensus the Royal Society of London, Series B 257:323–372. methods and their use in bootstrapping. Molecular Walker, A. D. 1972. New light on the origin of birds Biology and Evolution 13:437–444. and crocodiles. Nature 237:257–263. Wilkinson, M. 2003. Missing entries and multiple Walker, A. D. 1977. Evolution of the pelvis in birds and trees: Instability, relationships, and support in par­ dinosaurs. Pages 319–358 in Problems in Vertebrate simony analysis. Journal of Vertebrate Paleontol­ Evolution (S. M. Andrews, R. S. Miles, and A. D. ogy 23:311–323. Walker, Eds.). Linnean Society, London. Wilkinson, M., and J. L. Thorley. 2001. No compro­ Walker, A. D. 1980. The pelvis of Archaeopteryx. Geo­ mise on consensus. Taxon 50:181–184. logical Magazine 177:595–600. Witmer, L. M. 1990. The craniofacial system of Walker, A. D. 1985. The braincase of Archaeopteryx. Mesozoic birds (Aves). Zoological Journal of the Pages 123–134 in The Beginnings of Birds: Proceed­ Linnean Society 100:327–378. ings of the International Archaeopteryx Conference Witmer, L. M. 2002. The debate on avian ancestry: Phy­ Eichstatt 1984 (M. K. Hecht, J. H. Ostrom, G. Viohl, logeny, function, and fossils. Pages 3–30 in Meso­ and P. Wellnhofer, Eds.). Freunde des Jura-Muse­ zoic Birds: Above the Heads of Dinosaurs (L. M. ums Eichstatt, Eichstatt. Chiappe and L. M. Witmer, Eds.). University of Walker, A. D. 1990. A revision of Sphenosuchus acutus California Press, Berkeley. Haughton, a crocodylomorph reptile from the Wu, X., and S. Chatterjee. 1993. Dibrothrosuchus elaph- Elliot Formation (late Triassic or ) ros, a crocodylomorph from the Lower Jurassic of of South Africa. Philosophical Transactions of the China and the phylogeny of the Sphenosuchia. Royal Society of London, Series B 330:1–120. Journal of Vertebrate Paleontology 13:58–89. Wang, X., Z. Zhou, F. Zhang, and X. Xu. 2002. A nearly Xu, X. 2006. Feathered dinosaurs from China and the complete articulated rhamphorhynchoid pterosaur evolution of major avian characters. Integrative with exceptionally well-preserved wing mem­ Zoology 1:4–11. branes and “hairs” from , northeast Xu, X., Y. Cheng, X. Wang, C. Chang, and H. Chang. China. Chinese Science Bulletin 47:226–232. 2002a. An unusual oviraptorosaurian dinosaur Weishampel, D. B., P. Dodson, and H. Osmólska, Eds. from China. Nature 419:291–293. 2004. The Dinosauria, 2nd ed. University of Califor­ Xu, X., J. M. Clark, C. A. Forster, M. A. Norell, G. M. nia Press, Berkeley. Erickson, D. A. Eberth, C. Jia, and Q. Zhao. 2006. Welles, S. P. 1984. Dilophosaurus wetherilli (Dinosauria: A basal tyrannosauroid dinosaur from the Late Theropoda). Osteology and comparisons. Palaeon­ Jurassic of China. Nature 439:715–718. tographica A 185:85–180. Xu, X., and M. A. Norell. 2004. A new troodontid Welles, S. P., and R. A. Long. 1974. The tarsus of dinosaur from China with avian-like sleeping pos­ theropod dinosaurs. Annals of the South African ture. Nature 431:838–841. Museum 64:191–218. Xu, X., M. A. Norell, X. Kuang, X. Wang, Q. Zhao, Wellnhofer, P. 2004. The plumage of Archaeopteryx: and C. Jia. 2004. Basal tyrannosauroids from China Feathers of a dinosaur? Pages 282–300 in Feathered and evidence for protofeathers in tyrannosauroids. Dragons: Studies on the Transition from Dinosaurs Nature 431:680–684. to Birds (P. J. Currie, E. B. Koppelhus, M. A. Shugar, Xu, X., M. A. Norell, X. Wang, P. J. Makovicky, and X. and J. L. Wright, Eds.). Indiana University Press, Wu. 2002b. A basal troodontid from the Early Cre­ Bloomington. taceous of China. Nature 415:780–784. Welman, J. 1995. Euparkeria and the origin of birds. Xu, X., Z. Tang, and X. Wang. 1999a. A therizinosau­ South African Journal of Science 91:533–537. roid dinosaur with integumentary structures from Welten, M. C. M., F. J. Verbeek, A. H. Meijer, and China. Nature 399:350–354. M. K. Richardson. 2005. Gene expression and Xu, X., X. Wang, and X. Wu. 1999b. A dromaeosaurid digit homology in the chicken embryo wing. Evo­ dinosaur with a filamentous integument from the lution & Development 7:18–28. Yixian Formation of China. Nature 401: 262–266. Whetstone, K. N., and L. D. Martin. 1979. New Xu, X., and X. Wu. 2001. Cranial morphology of Sinor- look at the origin of birds and crocodiles. Nature nithosaurus millenii Xu et al. 1999 (Dinosauria: 279:234–236. Theropoda: Dromaeosauridae) from the Yixian Whetstone, K. N., and P. J. Whybrow. 1983. A “curso­ Formation of Liaoning, China. Canadian Journal of rial” crocodilian from the Triassic of Lesotho (Basuto­ Earth Sciences 38:1739–1752. land), southern Africa. University of Kansas Museum Xu, X., Z. Zhou, and X. Wang. 2000. The smallest known of Natural History, Occasional Papers no. 106. non-avian theropod dinosaur. Nature 408:705–708.

OM66_Text.indd 50 3/31/09 5:55:15 PM origin of birds 51

Xu, X., Z. Zhou, X. Wang, X. Kuang, F. Zhang, and Zhou, Z. 2004. The origin and early evolution of birds: X. Du. 2003. Four-winged dinosaurs from China. Discoveries, disputes, and perspectives from fossil Nature 421:335–340. evidence. Naturwissenschaften 91:455–471. Xu, X., Z. Zhou, X. Wang, X. Kuang, F. Zhang, and Zhou, Z., and L. Hou. 2002. The discovery and study X. Du. 2005. Origin of flight—Could “four-winged” of Mesozoic birds in China. Pages 160–183 in Meso­ dinosaurs fly? (Reply). Nature 438:E3–E4. zoic Birds: Above the Heads of Dinosaurs (L. M. Yates, A. M., and C. C. Vasconcelos. 2005. Furcula- Chiappe and L. M. Witmer, Eds.). University of like clavicles in the prosauropod dinosaur Mas- California Press, Berkeley. sospondylus. Journal of Vertebrate Paleontology Zhou, Z., and L. Martin. 1999. Feathered dinosaur 25:466–468. or bird? A new look at the hand of Archaeopteryx. You, H., M. C. Lamanna, J. D. Harris, L. M. Chiappe, Smithsonian Contributions to Paleobiology 89:289– J. O’Connor, S. Ji, J. Lü, C. Yuan, D. Li, X. Zhang, 293. and others. 2006. A nearly modern amphibious Zhou, Z., X. Wang, F. Zhang, and X. Xu. 2000. bird from the Early Cretaceous of northwestern Im­portant features of Caudipteryx—Evidence from China. Science 312:1640–1643. two nearly complete new specimens. Vertebrata Zanno, L. E. 2006. The pectoral girdle and forelimb of PalAsiatica 38:241–254. the primitive therizinosauroid Falcarius utahensis Zhou, Z., and F. Zhang. 2001. Two new ornithurine (Theropoda, Maniraptora): Analyzing evolution­ birds from the Early Cretaceous of western Liaon­ ary trends within Therizinosauroidea. Journal of ing, China. Chinese Science Bulletin 46:1258–1264. Vertebrate Paleontology 26:636–650. Zhou, Z., and F. Zhang. 2002. A long-tailed, seed- Zhang, F., and Z. Zhou. 2000. A primitive enantio­ eating bird from the Early Cretaceous of China. rnithine bird and the origin of feathers. Science Nature 418:405–409. 290:1955–1959. Zhou, Z., and F. Zhang. 2003a. Anatomy of the primi­ Zhang, F., Z. Zhou, X. Xu, X. Wang, and C. Sulli- tive bird Sapeornis chaoyangensis from the Early Cre­ van. 2008. A bizarre Jurassic maniraptoran from taceous of Liaoning, China. Canadian Journal of China with elongate ribbon-like feathers. Nature Earth Sciences 40:731–747. 455:1105–1108. Zhou, Z., and F. Zhang. 2003b. Jeholornis compared Zheng, X., Z. Zhang, and L. Hou. 2007. A new enantio­ to Archaeopteryx, with a new understanding of rnithine bird with four long rectrices from the Early the earliest avian evolution. Naturwissenschaften Cretaceous of northern Heibei, China. Acta Geo­ 90:220–225. logica Sinica 81:703–708. Zhou, Z., and F. Zhang. 2005. Discovery of an ornithu­ Zhou, H., and L. Niswander. 1996. Requirement for rine bird and its implication for Early Cretaceous BMP signaling in interdigital apoptosis and scale avian radiation. Proceedings of the National Acad­ formation. Science 272:738–741. emy of Sciences USA 102:18998–19002.

OM66_Text.indd 51 3/31/09 5:55:15 PM 52 ORNITHOLOGICAL MONOGRAPHS NO. 66

Appendix 1: Specimens Examined

Institution Specimen Specimen number

UKMNH Archaeopteryx KU25350, cast of JM 2257 UKMNH Archaeopteryx Cast of BMNH 37001 UKMNH Archaeopteryx Cast of HMN 1880 UKMNH Bambiraptor AMNH 001 (AMNH FR: 30556) UKMNH Bambiraptor KUVP 129737 UKMNH Confuciusornis Cast of IVPP V 11552 UKMNH Confuciusornis Cast of IVPP V 51556 UKMNH Patagopteryx Cast of MACN-N-11 UKMNH Sinornis (Cathayornis) Epoxy cast of BPV 538a UKMNH Sinornis (Cathayornis) Epoxy cast of BPV 538b UKMNH Sinornis (Cathayornis) Epoxy cast of IVPP V 9769A UKMNH Unenlagia Cast of MCH PVPH 78 RTMP RTMP 88.121.39 RTMP Struthiomimus TMP 90.26.01 RTMP Troodon TMP 82.19.23 RTMP Troodon RTMP 86.36.457

Abbreviations: UKMNH = University of Kansas Museum of Natural History, RTMP = Royal Tyrrell Museum of Paleontology.

Appendix 2: Character List for 242 3. : absent (0); present (1). Chiappe (2002a). Characters, Notes, and References 4. Vaned integumentary structures on hind limb: not present or poorly developed (0); Except for the characters that have ordered well developed, extensive (1). states, character states are not polarized; that 5. dorsal body osteoderms: absent (0); present is, zero does not necessarily represent the ple­ (1). Benton (1999). siomorphic state. “?” indicates missing data for Cranium a taxon; “–” indicates a character state that was 6. Premaxilla, posterior process: short and not applicable for a taxon (e.g., dental characters blunt (0); elongate and extending along for Citipati). Characters that are common to most nasal-maxillary suture posterior to external analyses or that, to our knowledge, are unique to nares (1). Clark et al. (2002). the present analysis are not followed by source 7. Premaxillae: unfused (0); fused (1). Lü et al. citations. Characters that are not common to most (2002). analyses or that were taken from a specific study 8. Premaxilla: maxillary process contacts nasal are followed by citations of the sources from to form posterior border of nares (0); maxil­ which they were taken. If we altered the wording lary process reduced so that maxilla partici­ or structure of a character, we noted that the char­ pates broadly in external naris (1). Modified acter was modified when citing its source. Twenty- from Clark et al. (2002). one characters of the carpus, manus, palate, and 9. Premaxilla: crenulate margin on buccal edge basipterygoid; some characters of the palate; and absent (0); present (1). Clark et al. (2002). some characters of the tarsus were included but 10. internarial bar: rounded (0); flat (1). Clark et only turned on in the alternative analysis of the al. (2002). data matrix. These 21 characters are listed at the 11. Maxilla, reduced in lateral aspect: absent end (characters 222–242). (0); present (1). Martin and Czerkas (2000). 12. external naris: short with posterior margin Integument not reaching anterior border of antorbital 1. elongate integumentary structures wider at fossa (0); elongate with posterior margin tip than base, with follicle and tubular shaft reaching or overlapping anterior border or calamus: absent (0); present (1). of the antorbital fossa (1); elongate, tall, 2. Vaned integumentary structures: symmetric nearly vertical (2). Modified from Clark et (0); asymmetric (1). Clark et al. (2002). al. (2002).

OM66_Text.indd 52 3/31/09 5:55:15 PM origin of birds 53

13. interfenestral bar: not present (0); present temporal bar is dorsally concave (1). Clark (1). Makovicky et al. (2003). et al. (2002). 14. Antorbital fenestra: present and without 28. Postorbital bar: jugal and postorbital con­ any significant antorbital fossa (0); ­antor tributing equally to postorbital bar (0); bital fenestra and fossa well developed ascending process of jugal reduced and (1); antorbital fenestra and fossa absent (2). descending process of postorbital ventrally Modified from Benton (2004). elongate (1); absent (2). Ordered. Modified 15. Antorbital fossa, pronounced accessory from Clark et al. (2002). fenestra: absent (0); present (1). Modified 29. Postorbital bar: parallels quadrate, lower from Clark et al. (2002). temporal fenestra rectangular in shape (0); 16. Antorbital fossa, tertiary fenestra: absent jugal and postorbital approach or contact (0); present (1). Modified from Clark et al. quadratojugal to constrict lower temporal (2002). fenestra (1). Clark et al. (2002). 17. Narial region pneumaticity: apneumatic or 30. Postorbital/squamosal temporal bar an­ poorly pneumatized (0); extensively pneu­ teroposteriorly short, with subtriangular matized (1). Modified from Clark et al. laterotemporal fenestra: absent (0); present (2002). (1); no distinct laterotemporal fenestra pres­ 18. Orbit, shape: round in lateral or dorsolateral ent (2). Sereno (1991). view (0); dorsoventrally elongate (1); an­ 31. Jugal bar: large (0); thin, reduced in height, teroposteriorly elongate (2). Modified from strap like or rod like (1). Modified from Lü Clark et al. (2002). et al. (2002). 19. Orbit, size: not significantly enlarged rela­ 32. Quadratojugal: without a horizontal process tive to the rest of the skull, frontals not caudal to the ascending process (reversed arched (0); clearly inflated, frontals strongly “L” or “L-shape”) (0); or with process (i.e., arched (1). inverted “T” or “Y-shape”) (1); greatly re­ 20. Prefrontal: not expanded within orbit (0); duced (2); fused to the jugal and not distin­ expanded within orbit, meeting interor­ guishable (3). Ordered. Modified from Clark bital septum (1); absent (2). Modified from et al. (2002). Walker (1990) and Benton (1999). 33. Quadratojugal or quadratojugal/jugal: su­ 21. Prefrontal, size: broad dorsal exposure simi­ tured to quadrate (0); ligamentous (1). Mod­ lar to that of lacrimal (0); greatly reduced in ified from Ruben and Jones (2000). size (1). Modified from Clark et al. (2002). 34. Quadrate: without lateral, round cotyla on 22. lacrimal: caudal process absent or small the mandibular process (0); with cotyla (1). (inverted “L-shape”) (0); present and lacri­ Modified from Chiappe (2002a). mal “T-shaped” in lateral view (1); curved, 35. Quadrate, orbital process: absent (0); small opening caudally (reversed “C-shape”) (2). or broad, not distinct and pointed (1); Modified from Clark et al. (2002) and Lü distinct and pointed (2). Modified from et al. (2002). Chiappe (2002a). 23. lacrimal: supraorbital crests absent in adult 36. Quadrate, lateral border: straight or without (0); crest(s) or lateral expansions dorsal enlarged quadrate foramen (“paraquadratic and/or anterolateral to orbit (1). Modified foramen”) (0); with lateral tab that touches from Clark et al. (2002). squamosal and quadratojugal above an en­ 24. lacrimal: no enlarged foramen or foramina larged quadrate foramen (“paraquadratic opening laterally at the angle of the lacrimal foramen”) (1). Clark et al. (2002). (0); enlarged foramen or foramina open­ 37. Quadrate: apneumatic (0); pneumatic (1). ing laterally at the angle of the lacrimal (1). Witmer (1990). Clark et al. (2002). 38. Quadrate, articulation: singular with squa­ 25. Postfrontal: present (0); absent (1). Benton mosal or squamosal and opisthotic (0); bipar­ (1999). tite with dermal and endochondral bones, 26. Postorbital: present (0); absent (1). anteriorly with prootic, squamosal, and lat­ 27. Postorbital, lateral view: straight ante­ erosphenoid and posteriorly with prootic and rior (frontal) process (0); frontal process the fused opisthotic and exoccipital (1). Modi­ curves anterodorsally, and dorsal border of fied from Walker (1972) and Martin (1983).

OM66_Text.indd 53 3/31/09 5:55:15 PM 54 ORNITHOLOGICAL MONOGRAPHS NO. 66

39. Quadrate, head: single (0); double (1). 55. Maxilla, palatal shelf: flat (0); with midline Walker (1972) and Martin (1983). ventral tooth-like projection (1). Clark et al. 40. Quadrate: vertical (0); strongly inclined pos­ (2002). teroventrally so that the distal end lies far 56. Posterior maxillary sinus cup shaped: ab­ forward of the proximal end (1); strongly sent (0); present (1). Modified from Chiappe inclined anteriorly so that the proximal end (2002a). lies far forward of the distal end (2). Modi­ 57. ectopterygoid: present with no jugal “hook” fied from Walker (1972) and Clark et al. (0); attached to jugal by a distinctly “hook- (2002). like” process (1); present but not attached 41. Quadrate, distal end: one or two transversely to jugal (2); absent (3). Modified from El­ aligned condyles (0); triangular condylar pat­ zanowski (1999). , usually composed of three condyles (1). 58. ectopterygoid, position: ventral to or level Modified from Chiappe (2002a). with transverse flange of pterygoid (0); 42. Squamosal: with descending process con­ dorsal to transverse flange of pterygoid (1). tacting quadratojugal (0); with descending Modified from Sereno and Novas (1993). process not contacting quadratojugal (1); 59. ectopterygoid: no fossa on ventral surface without descending process (2). Modified (0); fossa on ventral surface (1). Modified from Clark et al. (2002). from Clark et al. (2002). 43. Squamosal, descending process: parallels 60. ectopterygoid: no fossa on dorsal surface (0); quadrate shaft (0); nearly perpendicular to fossa on dorsal surface (1). Clark et al. (2002). quadrate shaft (1). Clark et al. (2002). Braincase and otic region 44. Squamosal: distinct and separate element 61. Occipital condyle: larger than or equal to (0); incorporated into braincase (1). Modi­ foramen magnum in size (0); smaller than fied from Chiappe (2002a). foramen magnum (1). 45. frontals: separate (0); fused (1). Holtz (2000). 62. Post-temporal fenestra: large (0); reduced 46. frontals: narrow anteriorly as a wedge be­ to a foramen or slit, or absent (1). Modified tween nasals (0); end abruptly anteriorly, from Sereno and Novas (1993). suture with nasal transversely oriented (1). 63. foramina for entrance of cerebral branches Clark et al. (2002). of internal carotid arteries into braincase: 47. frontals: anterior emargination of supratem­ positioned on the posteroventral surface of poral fossa straight or slightly curved (0); the parabasisphenoid (0); positioned on the strongly sinusoidal and reaching onto pos­ lateral surface or anterolateral surface of the torbital process (1). Currie (1995) and Clark parabasisphenoid (1). Gower (2002). et al. (2002). 64. Basisphenoidal recess: present (0); shallow 48. frontals, postorbital process in dorsal view: or absent (1). Currie and Zhao (1993b). smooth transition from orbital margin (0); 65. Parabasisphenoid, “semilunar depression”: sharply demarcated from orbital margin (1). present (0); absent (1). Gower and Weber Currie (1995) and Clark et al. (2002). (1998). 49. frontal: edge smooth in region of lacrimal 66. Cultriform process of parabasisphenoid: suture (0); edge notched (1). Currie (1995) base not highly pneumatized or pneu­ and Clark et al. (2002). matized but without an inflated bulla (0); 50. Parietals: separate (0); fused (1). Clark et al. highly pneumatized with base expanded (2002). and pneumatic, forming a parasphenoid 51. Skull, roof: not strongly convex (0); strongly bulla (1). Modified from Clark et al. (2002). convex (1). 67. Paroccipital process: straight, projects later­ Palate ally or posterolaterally (0); distal end curves 52. Vomers: not fused (0); fused anteriorly (1). ventrally, pendant (1). Clark et al. (2002). Gauthier (1986). 68. Paroccipital process: straight distal end (0); 53. Palatal teeth: present (0); absent (1). Benton distal end twisted to face posterodorsally (1999). (1). Currie (1995). 54. “Secondary palate”: no distinct “secondary 69. eustachian tubes: not enclosed by bone (0); palate” (0); “secondary palate” short (1); partially or fully enclosed (1). Gower and “secondary palate” extensive (2). Weber (1998) and Gower (2002).

OM66_Text.indd 54 3/31/09 5:55:16 PM origin of birds 55

70. Vestibule, ossification of medial wall: in­ 84. Premaxillary tooth crowns: in cross section complete (0); almost or completely ossified crowns suboval to subcircular (0); asym­ (1). Gower and Weber (1998) and Gower metrical (D-shaped in cross section) with (2002). flat lingual surface (1). Clark et al. (2002). 71. Metotic fissure: persisting as an undivided 85. Maxillary teeth: serrated (0); some without opening (i.e., the metotic foramen) (0); sub­ serrations anteriorly (1); some or all teeth divided during development with a foramen without serrations (2). Modified from Clark for the vagus (X) nerve and jugal vein (“va­ et al. (2002). gal” or “jugular foramen”) separated via an 86. dentary teeth: serrated (0); some without osseous, “prevagal commissure” from the serrations anteriorly (1); some or all teeth lateral aperture (= fenestra pseudorotunda) without serrations (2). Modified from Clark of a recessus scalae tympani (1). Whetstone et al. (2002). and Martin (1979), Rieppel (1985), Gower 87. dentary and maxillary teeth: less than 25 in and Weber (1998).1 dentary (0); 25–30 in dentary (1); teeth rela­ 72. Perilymphatic foramen, position: medial tively small, and numerous (more than 30 and oriented so as to transmit the peri­ in dentary) (2). Modified from Clark et al. lymphatic duct out of the otic capsule in a (2002). posteromedial or posterior direction (0); fo­ 88. dentary teeth, spacing: even (0); anterior ramen positioned more laterally so that the dentary teeth smaller, more numerous, and perilymphatic duct is transmitted postero­ more closely appressed than those in middle laterally or laterally and the foramen is at of tooth row (1). Clark et al. (2002). least partly visible in lateral view (1). Gower 89. dentary tooth implantation: in individual and Weber (1998), Gower (2002). sockets (0); some teeth in communal groove 73. Cochlear recess: not clearly differentiated or beginning posteriorly (1); all teeth in a com­ short (0); clearly differentiated and elongate munal groove (2). Ordered. Modified from (1). Walker (1990), Gower and Weber (1998), Chiappe (2002a). Gower (2002). 90. Serration denticles: not apically hooked (0); 74. hypoglossal (XII) nerve has three or more apically hooked (1). Modified from Clark et external foramina in posterolateral region of al. (2002) and from Makovicky and Norell braincase floor: absent (0); present (1). Cur­ (2004). rie and Zhao (1993b). 91. Teeth with expanded roots: absent (0); pres­ 75. Vagus (X) nerve diverted, exiting on the ent in at least some teeth (1); present in all occiput: absent (0); present (1). Currie and teeth (2). Martin et al. (1980) and Martin and Zhao (1993b). Stewart (1999). 76. Posterior tympanic recess: absent (0); opens 92. Premaxillary teeth with constriction be­ on anterior margin of paroccipital process tween crown and root: absent (0); present (1); opens within columellar recess (2). Or­ (1). Martin et al. (1980) and Martin and dered. Modified from Clark et al. (2002). Stewart (1999). 77. dorsal tympanic recess: absent (0); present 93. Maxillary teeth with constriction between (1). Witmer (1990). crown and root: absent (0); present (1). 78. Anterior tympanic recess: absent (0); pres­ Martin et al. (1980) and Martin and Stewart ent (1). Witmer (1990). (1999). 79. Contralateral communication between at 94. dentary teeth with constriction between least some tympanic diverticulae: absent (0); crown and root: absent (0); present (1). Martin present (1). et al. (1980) and Martin and Stewart (1999). Dentition2 95. Oval resorption pit on lingual aspect of root 80. Premaxilla: toothed (0); toothless (1). surrounding developing crown and closed 81. Maxilla: toothed (0); toothless (1). at bottom: absent (0); present (1). Martin et 82. dentary: toothed (0); toothless (1). al. (1980) and Martin and Stewart (1999). 83. Premaxillary teeth: serrated (0); some with­ out serrations anteriorly (1); some or all 96. interdental plates: present (0); absent (1). teeth without serrations (2). Modified from Martin et al. (1980) and Martin and Stewart Clark et al. (2002). (1999).

OM66_Text.indd 55 3/31/09 5:55:16 PM 56 ORNITHOLOGICAL MONOGRAPHS NO. 66

97. Mandibular symphysis: not ossified or only medial process or both (1). Modified from loosely sutured (0); tightly sutured (1); fused Elzanowski (1999). (2). Ordered. Modified from Marya´nska et 112. Articular pneumaticity: absent (0); present al. (2002). (1). Witmer (1990). 98. dentary: symphyseal region broad and 113. retroarticular process: absent (0); short, straight paralleling lateral margin (0); medi­ stout (1); elongate and slender (2); elongate ally recurved slightly (1); strongly recurved and slender, with vertical columnar process medially (2). Clark et al. (2002). rising from posteromedial corner (3). Modi­ 99. dentary: symphyseal region in line with fied from Currie (1995) and from Clark et al. main part of buccal edge (0); symphyseal (2002). edge downturned (1). Clark et al. (2002). Axial skeleton 100. dentary, posterior end: without posterodor­ 114. Cervicodorsal hypapophyses: absent (0); sal process dorsal to mandibular fenestra present (1). or with only a small posterodorsal process 115. Axial and postaxial epipophyses: absent or over anterior part of the external mandibu­ poorly developed, not extending past pos­ lar fenestra (0); well-developed postero­ terior rim of postzygapophyses (0); large, dorsal process extending over most of the posteriorly directed, extended beyond external mandibular fenestra (1). Modified postzygapophyses (1). Modified from Clark from Clark et al. (2002). et al. (2002). 101. dentary, labial face: flat (0); with lateral ridge 116. Axial neural spine: flared transversely (0); and inset tooth row (1). Clark et al. (2002). compressed mediolaterally (1). Clark et al. 102. dentary, nutrient foramina on external sur­ (2002). face: superficial (0); lying in a deep groove 117. Axial neural spine: does not extend anteri­ (1). Clark et al. (2002). orly beyond prezygapophyses (0); extends 103. intramandibular joint: absent or poorly de­ anteriorly beyond prezygapophyses (1). veloped (0); present and well developed (1). Tykoski and Rowe (2004). Gauthier (1986) and Benton (1999). 118. Postaxial intercentra: present (0); absent (1). 104. external mandibular fenestra: present (0), Sereno (1991). absent (1). 119. Anterior cervical centra: subcircular or square 105. external mandibular fenestra: not subdi­ in anterior view (0); distinctly wider than vided by a spinous anterior process of the high, kidney-shaped (1). Clark et al. (2002). surangular (0); subdivided by a spinous an­ 120. Posterior cervical vertebrae: no carotid terior process of the surangular (1). Clark et processes (0); carotid processes present (1). al. (2002). Clark et al. (2002). 106. Surangular, lateral surface: no foramen for 121. Cervical vertebrae, neural spines: antero­ process of dentary (0); foramen in lateral posteriorly long (0); short and centered on surface for process of dentary (1). Modified neural arch (1). Clark et al. (2002). from Clark et al. (2002). 122. Cervical and anterior trunk vertebrae: pro­ 107. Surangular, lateral surface: no horizontal coelus (0); amphicoelus (1); opisthocoelus shelf (0); prominent, laterally expanding (2); incipiently heterocoelus (3); fully hetero­ horizontal shelf present anteroventral to the coelus (4); platycoelus (5); anterior articular mandibular condyle (1). Holtz (2004). surface flat, posterior surface weakly con­ 108. Splenial: not widely exposed on lateral sur­ cave (6). Modified from Clark et al. (2002). face of mandible (0); exposed as a broad tri­ 123. dorsal vertebrae, hyposphene–hypantrum angle between dentary and angular on lateral accessory intervertebral articulations: ab­ surface of mandible (1). Clark et al. (2002). sent (0); present (1). Gauthier (1986). 109. Coronoid ossification: present (0); vestigial 124. dorsal vertebrae, transverse processes: or absent (1). Elzanowski (1999). straight or slightly backswept (0); strongly 110. Articular and surangular coossified: absent posteriorly backswept, triangular in dorsal (0); present (1). Elzanowski (1999). view (1). Tykoski and Rowe (2004). 111. Articular, surface for quadrate: with nei­ 125. dorsal vertebrae: zygapophyses abut ther lateral nor medial processes (0); with one another above neural canal, oppo­ development of either a lateral process or a site hyposphenes meet to form lamina (0);

OM66_Text.indd 56 3/31/09 5:55:16 PM origin of birds 57

zygapophyses placed lateral to neural canal 140. Sternal ribs: absent or cartilaginous (0); ossi­ and separated by groove for interspinous fied (1). Modified from Clark et al. (2002). ligaments, hyposphenes separated (1). Clark 141. Sternum: absent or cartilaginous (0); ossified et al. (2002). plates separate in adult (1); ossified plates 126. Caudal dorsal vertebrae, parapophyses: fused in adult (2). Ordered. Modified from flush with neural arch (0); distinctly pro­ Clark et al. (2002). jected on pedicels (“stalked”) (1). Modified 142. Sternum, : absent or weakly developed from Clark et al. (2002). (0); present and limited to posterior margin 127. dorsal centra, pleurocoels: absent (0); present of sternum (1); extends to anterior margin on some centra (1); present on all centra (2). of sternum (2). Ordered. Modified from Ordered. Modified from Clark et al. (2002). Chiappe (2002a). 128. dorsal vertebrae neural spines: scars for in­ 143. Sternum, costal facets: absent (0); present terspinous ligaments terminate at apex of (1). Chiappe (2002a). neural spine (0); terminate below apex of 144. Sternum, lateral processes: absent (0); pres­ neural spine (1). Clark et al. (2002). ent (1). Modified from Clark et al. (2002) and 129. Presacral vertebrae: apneumatic (0); cervi­ Chiappe (2002a). cals or dorsals pneumatic (1); all presacral 145. Sternum, anterior margin grooved for re­ vertebrae pneumatic (2). Modified from ception of coracoids: absent (0); present (1). Clark et al. (2002). Clark et al. (2002). 130. Sacral vertebrae, number: two (0); three– 146. Sternum, articular facets for coracoid: lat­ four (1); five (2); six (3); seven or more (4). eral or anterolateral (0); almost or fully ante­ Benton (1999) and Clark et al. (2002). rior (1). Modified from Clark et al. (2002). 131. Synsacrum: absent in adults (0); present in Appendicular skeleton (pectoral girdle and fore­ adults (1). Holtz (2000). limb) 132. Caudal vertebrae: 40 or more (0); 39–25 (1); 147. interclavicle: present (0); rudimentary or fewer than 25 (2). Ordered. Modified from absent (1); incorporated into sternum (2); Clark et al. (2002). incorporated into furcula (3). Modified from 133. Caudal vertebrae: caudals homogenous Benton (1999) and Martin et al. (1998a).3 in shape, without transition point (0); dis­ 148. Clavicles: present (0); rudimentary or absent tinct transition point in caudal series, from (1); fused into furcula (2). Modified from shorter centra with long transverse pro­ Benton (1999).4 cesses proximally to longer centra with 149. furcula, shape: broad “V-shaped” (0); boo­ small or no transverse processes distally (1). merang shaped (1); narrow “V-shaped” (2); Modified from Clark et al. (2002). “U-shaped” (3). Modified from Chiappe 134. Caudal vertebrae: transition point in caudal (2002a). series begins distal to the 10th caudal (0); at 150. furcula, hypocleidium: absent (0); small (1); or proximal to the 10th caudal (1). Clark et prominent, long (2). Clark et al. (2002). al. (2002). 151. Scapula, acromion process: absent (0) weakly 135. Caudal vertebrae: zygapophyses not elon­ developed (1); quadrangular (2); columnar gate (0); elongate, extending past centrum to or triangular, projecting cranially away from contact other centra (1); extremely elongate, corpus of scapula and glenoid fossa, but not ossified tendons well developed (2). Ordered. laterally everted in dorsal view (3); project­ Ostrom (1969) and Clark et al. (2002). ing cranially away from corpus of scapula 136. Pygostyle: absent (0); present (1). Modified and glenoid fossa and laterally everted in from Chiappe (2002a). dorsal view (4). 137. Cervical ribs: shafts slender and each lon­ 152. Scapula, shape: not elongate and strap- ger than to which it articulates (0); like (0), elongate and strap-like (1). Martin broad and shorter than vertebra (1); fused (2004). (2). Clark et al. (2002). 153. Scapula, shaft: straight (0); sagittally curved 138. gastralia: present (0); absent (1). (1). Chiappe (2002a). 139. Uncinate processes: absent or cartilaginous 154. Scapula, position: not parallel to the ver­ (0); ossified (1). Modified from Clark et al. tebral column (0); parallel to the vertebral (2002). column (1). Martin (2004).

OM66_Text.indd 57 3/31/09 5:55:17 PM 58 ORNITHOLOGICAL MONOGRAPHS NO. 66

155. Scapula and coracoid: articulate through a somewhat asymmetrical, with the radial wide suture or fused (0); meeting at a mobile condyle noticeably larger and taller than joint (1). Modified from Clark et al. (2002). the ulnar condyle (1); strongly asymmetri­ 156. Scapula and coracoid: angle formed between cal with the radial condyle much larger and scapula and coracoid obtuse (0); coracoid taller than the ulnar condyle (2). Paul (2002) reflected, forming right angle to the scapula and Gishlick (2001). (1); coracoid reflected at an acute angle to 171. humerus: without well-developed brachial the scapula (2). Ordered. depression on anterior face of distal hu­ 157. Coracoid: short (0), elongated with trap­ merus (0); well-developed brachial depres­ ezoidal profile (1); strut-like (2). Ordered. sion present (1). Chiappe (2002a). Chiappe (2002a). 172. humerus: without well-developed olecra­ 158. Coracoid: no proximal constriction to form non fossa on posterior face of distal end “neck” (0); proximally constricted (1). Mod­ of humerus (0); well-developed olecranon ified from Senter et al. (2004). fossa present (1). Chiappe (2002a). 159. Coracoid in lateral view: not distinctly tri­ 173. Ulna, proximal articular surface: single con­ angular (0); distinctly triangular (1). tinuous articular facet (0); divided into at 160. Coracoid, acrocoracoid process: not present least two distinct fossae separated by a me­ or poorly developed (0), present (1). Modi­ dian ridge (1). Clark et al. (2002). fied from Chiappe (2002a). 174. Ulna: olecranon process weakly developed 161. Coracoid, procoracoidal process: not pres­ (0); distinct and large (1). Clark et al. (2002). ent or poorly developed (0); present (1). 175. Ulna, distal articular surface: flat (0); slightly Chiappe (2002a). convex (1); strongly convex (2). Ordered. 162. Coracoid, lateral margin: not distinctly con­ Modified from Clark et al. (2002). vex (0); distinctly convex (1). Chiappe and 176. Ulna, distal surface: labrum condyli not well Walker (2002). developed (0); well developed (1). Modified 163. Coracoid, supracoracoid nerve foramen: from Chiappe (2002a). superficial (0); opens into an elongate­ fur Appendicular skeleton (pelvic girdle and hind row medially, separated from medial mar­ limb) gin of the coracoid by a thick, bony bar (1). 177. ilium, anterior process: ventral edge straight Chiappe and Walker (2002). or gently curved (0); ventral edge hooked 164. Triosseal canal: not present (0); present (1). anteriorly (1); very strongly hooked (2). 165. glenoid fossa: lateral (0); posteroventral (1); Modified from Clark et al. (2002). dorsolateral (2). Martin (1983, 2004). 178. ilium, preacetabular process: shorter than 166. Proximal humerus: medial margin of humerus or nearly equal to postacetabular process of under the inner trochanter nearly straight, ilium (0); preacetabular process of ilium ex­ not strongly arched (0); medial margin of panded, longer than postacetabular process humerus strongly arched under a prominent of ilium (1). internal tuberosity, often approaching hori­ 179. ilium, postacetabular process, lateral view: zontal inclination (1). Sereno (1991). squared (0); somewhat convex (1); acumi­ 167. humerus: no distinct transverse ligamental nate (2). Ordered. Modified from Clark et al. groove (0); distinct transverse ligamental (2002). groove present (1). 180. ilium, brevis fossa: absent (0); shelflike (1); 168. humerus, bicipital crest: absent or rudimen­ deeply concave with lateral overhang (2). tary (0); present (1). Modified from Chiappe Modified from Clark et al. (2002). (2002a). 181. ilium, “cuppedicus” fossa: absent or poorly 169. humerus, deltopectoral crest: distinct, quad­ developed (0); present and formed as an an­ rangular, or triangular (0); less pronounced, tiliac shelf anterior to the acetabulum, extend­ forming an arc rather being quadrangular ing posteriorly to above anterior end of the (1); proximal humerus with rounded edges acetabulum (1); posterior end of fossa on ante­ and poorly developed deltopectoral crest rior end of pubic peduncle, anterior to acetab­ (2). Modified from Clark et al. (2002). ulum (2). Modified from Clark et al. (2002). 170. humerus, distal condyles: symmetrical, 182. ilium, “cuppedicus” fossa: deep fossa approximately equal in size and height (0); ventrally concave (0); shallow or slightly

OM66_Text.indd 58 3/31/09 5:55:17 PM origin of birds 59

concave, no lateral overhang (1). Modified 200. : lesser trochanter separated from from Clark et al. (2002). greater trochanter by deep cleft (0); tro­ 183. ilium, antitrochanter: absent (0); weakly de­ chanters separated by a small groove (1); veloped (1); prominent (2). Ordered. trochanters completely fused to form crista 184. ilium: supracetabular crest present (0); su­ trochanteris (2); greater and lesser trochant­ pracetabular crest absent (1). Naples et al. ers not well developed (3). Modified from (2002). Clark et al. (2002). 185. Acetabulum: imperforate (0); semiperfo­ 201. femur, lesser trochanter: absent or not rate (1); fully perforate (2). Ordered. Benton well developed (0); distinct but small, not (1999). elongate (1); alariform (2); cylindrical in 186. Pubes: vertical (0); propubic (1); opisthopu­ cross section (3). Modified from Clark et al. bic (2). Naples et al. (2002). (2002). 187. Pubes: shorter than or nearly equal in length 202. femur: posterior trochanter absent or repre­ to ischia (0); longer than ischia (1). Sereno sented by rugose area (0); distinctly raised (1991). from shaft and moundlike (1); forming 188. Pubes: fused in distal symphysis (0); pubic a distinct ridge (2). Some confusion sur­ bones separate (1). rounds the meaning of the term “posterior 189. Pubes: pubic shaft straight (0); distal end trochanter.” Here, the interpretation of Os­ curves anteriorly, anterior surface of shaft trom (1969) is followed rather than that of concave (1). Clark et al. (2002). Gauthier (1986). Clark et al. (2002). 190. Pubes: no strong posterior “kink” in pubes 203. femur, fourth trochanter: present (0); indis­ (0); strong posterior “kink” at midshaft (1). tinct or absent (1). Clark et al. (2002). Modified from Senter et al. (2004). 204. femur, anterior surface: no crest proximal to 191. Calceus pubis: absent (0); present and with medial distal condyle (0); crest proximal to anterior and posterior projections (1); with medial distal condyle extending onto ante­ little or no anterior projection (2). Modified rior surface of shaft (1). Clark et al. (2002). from Clark et al. (2002). 205. fibula: reaches the proximal tarsals (0); re­ 192. ischium, proximodorsal process approach­ duced distally and not meeting the proxi­ ing or abutting the ventral margin of the mal tarsals (1). ilium: absent (0); present (1). Chiappe 206. fibula, anterior trochanter: developed as a (2002a). low vertical crest or oval rugosity, or tro­ 193. ischium, obturator process: absent (0); pres­ chanter absent (0); robust, knob-shaped, ent and proximal in position (1); located giving fibula a crooked profile in lateral near middle of ischiadic shaft (2); located or medial view (1). Modified from Sereno at distal end of ischium (3). Modified from (1991). Clark et al. (2002). 207. Tibia, medial cnemial crest: absent (0); pres­ 194. ischium, obturator process: does not contact ent (1). Modified from Clark et al. (2002). pubis (0); contacts pubis (1); forms a broad 208. Tibia, calcaneum, and astragalus: unfused pubioischiadic plate (2). Clark et al. (2002). (0); partly fused (1); completely fused, form­ 195. ischium, obturator notch or foramen: pres­ ing a tibiotarsus (2). Ordered. Modified from ent (0); absent (1). Modified from Clark et al. Chiappe (2002a). (2002). 209. Supratendinal bridge: absent (0); present 196. ischium: ischial boot (expanded distal end) (1). Chiappe (2002a). absent (0); present (1). Modified from Clark 210. Astragalus, tibial articular surface flexed: et al. (2002). absent (0); present (1). Sereno (1991). 197. femur: proximal articular surface not ex­ 211. Calcaneal facets for fibula and distal tarsal tended under femoral head (0); extended four: separated (0); contiguous (1). Sereno under femoral head (1). Sereno and Arcucci (1991). (1994). 212. Calcaneal condyle hemicylindrical: absent 198. femoral head, orientation: oblique (0); at (0); present (1). Benton (1999). right angle to shaft (1). 213. Proximal tarsals and distal tarsals: meeting 199. femoral neck: absent (0); present (1). in a hinge, forming simple mesotarsal ankle Chiappe (2002a). (0); not forming a mesotarsal ankle (1).

OM66_Text.indd 59 3/31/09 5:55:17 PM 60 ORNITHOLOGICAL MONOGRAPHS NO. 66

214. distal tarsals: free (0); partially fused to 230. radiale: simple, discoid, or quadrangular metatarsals (1); completely fused to meta­ (0); elongate, rod-like (1); articular surface tarsals (2); fused and forming a “tarsal cap” sinuous, complex (2). Walker (1990), Parrish (3). Martin (1983, 1991, 2004). (1993). 215. hypotarsus: not present (0); present (1). 231. Carpometacarpus: absent (0); distal carpals Chiappe (2002a). and metacarpals partially fused (1); distal car­ 216. Metatarsals: unfused (0); proximodistal fu­ pals and metacarpals completely fused (2). sion forming tarsometatarsus (1); distoprox­ 232. Carpometacarpus, extensor process: absent imal fusion forming tarsometatarsus (2). or poorly developed (0); well developed (1). Martin (1991). 233. Manus: pentadactyl (0); digits I, II, III prom­ 217. Metatarsus configuration: metatarsals di­ inent, digits IV and V greatly reduced or verging from ankle (0); compact with meta­ absent (1); digits II, III, IV prominent, digits tarsals I–IV tightly bunched (1). Modified I and V greatly reduced or absent (2); didac­ from Benton (1999). tyl or with only two well-developed digits 218. Metatarsal III: visible between metatarsals II (3); reduced to a single digit (4). and IV (0); pinched between metatarsals II 234. Second (topographical) metacarpal: un­ and IV, the latter two contacting one another modified (0); more robust than other meta­ proximally in front of III (1); not reaching carpals (1). Modified from Zhou and Martin proximal end of metatarsus (2). Ordered. (1999). Modified from Clark et al. (2002). 235. Third (topographical) metacarpal: unmodi­ 219. Metatarsal V: present with “hooked” proxi­ fied (0); bowed (1); slender, appressed to mal end (0); present without “hooked” the second (topographical) metacarpal (2); proximal end (1); reduced to a splint (2); ab­ greatly reduced or absent (3). Modified sent (3). Ordered. from Gauthier (1986) and Zhou and Martin 220. Second pedal digit with hypertrophied (1999). claw: absent (0); present (1). 236. Third (topographical) metacarpal slants 221. hallux: not reversed (0); partially reversed ventrally towards distal end: absent (0); (1); completely reversed (2); hallux ab­ present (1). Modified from Zhou and Martin sent (3). Ordered. Modified from Chiappe (1999). (2002a). 237. Third (topographical) metacarpal: does not Characters Turned On for Alternative Analysis project distally more than second (topo­ 222. Basipterygoid process: well developed (0); graphical) metacarpal (0); projects distally abbreviated (1); absent (2). beyond second (topographical) metacarpal 223. Palatine: without elongate maxillary process (1). Chiappe and Walker (2002). (0); with elongate maxillary process (1). 238. Second (topographical) digit, first phalanx 224. Palatine: without “hook-shaped” process posterolaterally expanded: absent (0); pres­ enclosing choana (0); with process (1). ent (1). Modified from Zhou and Martin 225. Palatine: without broad pterygoid wing (0); (1999). with broad pterygoid wing (1). 239. Third (topographical) digit, phalanx count: 226. Palatine: tetraradiate with jugal process (0); four or more (0); one–four (1); zero (2). triradiate without jugal process (1). 240. Third (topographical) digit, proximal pha­ 227. Pterygoid: without basal process (0); with lanx: similar in size to or smaller than the basal process (1). second phalanx of the digit (0); more than or 228. Carpus: more than four free carpals (0); four equal to twice the length of the second pha­ free carpals (1); three or fewer free carpals lanx (1). Modified from Senter et al. (2004). (2). 241. Third (topographical) digit: without any 229. Ulnare: simple, discoid or quadrangular but characteristic “twist” (0); phalanges twisted not “V-shaped” (0); ulnare elongate, rod-like about long axis with distal articular con­ (1); replaced during development by a “V- dyles directed medially when in articulation shaped” element (the pisiform) (2); absent (1). Modified from Wagner and Gauthier and not replaced by a “V-shaped” element (1999) and Gishlick (2001). (3). Hinchliffe and Hecht (1984), Hinchliffe 242. Ascending process of astragalus: absent (0); (1985). present (1).

OM66_Text.indd 60 3/31/09 5:55:18 PM origin of birds 61

Notes the subdivided metotic foramen transmits cranial nerves X–XII and the internal jugular vein and 1 Although widely discussed in connection should be referred to as the “vagal” or “jugular with the origin of birds (e.g., Whetstone and foramen” (Gower and Weber 1998). These ves­ Martin 1979; Walker 1985, 1990; Chatterjee 1991; sels are shunted posteriorly toward the occiput, Gower and Weber 1998), a metotic fissure char­ and ultimately cranial nerves X–XII and the in­ acter has not, to our knowledge, been used in ternal jugular vein exit posteriorly on the oc­ cladistic analyses of the origin of birds. Termino­ ciput (Walker 1985, 1990; Gower and Weber 1998; logical confusion and conflicting accounts of the Gower and Walker 2002). In early evolutionary development of the metotic fissure have plagued stages, the jugular–vagal foramen may lie close these discussions, but Rieppel (1985) and more to the fenestra pseudorotunda, but it would still recently Gower and Weber (1998) offer excellent be distinct from this opening. reviews and figures. Briefly, the metotic fissure is The homology of the prevagal commissures a gap in the embryonic skull of reptiles and birds, subdividing the metotic fissure–foramen in birds between the basal plate of the basicranium and and crocodylians has been debated (e.g., de Beer the occipital arch on one side and the otic capsule and Barrington 1934; de Beer 1937; Whetstone on the other side (de Beer and Barrington 1934, and Martin 1979; Rieppel 1985; Walker 1985, 1990; Rieppel 1985). During development, this gap is Chatterjee 1991; Gower and Weber 1998), but the restricted by the growth of a basicapsular com­ balance of evidence indicates that these structures missure extending posteriorly from the root of are homologous (pro, de Beer and Barrington the facial (VII) nerve. The glossopharyngeal (IX), 1934; de Beer 1937; Whetstone and Martin 1979; vagus (X), and accessory (XI) nerves and a vein Walker 1985, 1990; contra, Rieppel 1985, Chatter­ (probably the internal jugular vein) traverse the jee 1991; see also the extensive review by Gower remaining portion of the gap. If this gap remains and Weber 1998). Gower and Weber (1998:398) undivided into maturity, the opening should be correctly observed that the homology of the avian referred to as the “metotic foramen” (Gower and and crocodylian structures has “perhaps prema­ Weber 1998). In the adult skull it is located pos­ turely been considered as (negatively) resolved teroventral to the fenestra ovalis (into which the in the wake of the currently favoured ‘theropod footplate of the stapes fits). hypothesis’ of avian origins.” In most archosaurs the metotic foramen Safely inferring the presence of a true fenestra remains undivided, but in crocodylomorphs, pseudorotunda requires evidence of a prevagal birds, and some theropods, the metotic foramen commissure subdividing the metotic foramen; is subdivided. In birds and crocodylian croco­ putative correlates (presence of a lateral ridge on dylomorphs, for which developmental data are the exoccipital, shift in the orientation of the peri­ available, an osseous “prevagal commissure” de­ lymphatic foramen, etc.) are insufficient. A clear velops that separates the metotic fissure–foramen structure that can be referred to as the “fenes­ into anterior and posterior parts (Rieppel 1985, tra pseudorotunda” must exist for the derived Gower and Weber 1998, Gower and Walker 2002). condition to be considered present (Gower and The anterior part of the subdivided metotic fora­ Weber 1998). We used this criterion when formu­ men, housing the perilymphatic sac, is properly lating and coding a metotic fissure character for referred to as the “recessus scalae tympani” (Riep­ inclusion in our analysis. For example, we coded pel 1985, Gower and Weber 1998). The route of Dibrothosuchus as having a subdivided metotic the perilymphatic duct is shifted during the sub­ fissure because such a structure is clearly visible division of the metotic fissure–foramen such that (Wu and Chatterjee 1993), whereas Dromaeosau- through the lateral aperture of the recessus scalae rus, in which a fenestra pseudorotunda is merely tympani, the perilymphatic sac fuses with the inferred to be present on the basis of putative tympanic mucous membrane, forming a second­ correlates (Currie 1995), was not coded with the ary tympanic membrane (Whetstone and Martin derived state because the metotic fissure was 1979, Gower and Weber 1998). This membrane clearly not subdivided (state 0) and was only as­ stretches across the osseous frame of the lateral sumed to be subdivided. The same criterion was aperture of the recessus scalae tympani; together used in coding the metotic fissure as subdivided they are properly referred to as the “fenestra in the other 11 taxa (including some birds, other pseudorotunda.” The more posterior aperture of crocodylomorphs, and some maniraptorans) for

OM66_Text.indd 61 3/31/09 5:55:18 PM 62 ORNITHOLOGICAL MONOGRAPHS NO. 66

which the derived state is listed as present in our of the origin of birds, Sereno (1991) used the matrix. absence of an interclavicle as a synapomorphy 2 Dental morphology and patterns of tooth of the Ornithodira, which he defined as “Ptero­ implantation in birds and other archosaurs have sauria, Scleromochlus, Dinosauromorpha (includ­ been widely discussed since Martin et al.’s (1980) ing birds), and all descendants of their common demonstration that Mesozoic birds, including Ar- ancestor” (p. 34). Padian (2001b:492) went fur­ chaeopteryx, share with crocodyliform crocodylo­ ther and claimed that the interclavicle is “lost or morphs almost identical dental morphologies unknown in most archosaurs.” Padian (2001b) and implantation systems (see also Martin 1983, follows Gauthier (1986) in restricting the name 1985, 1991; Martin and Stewart 1999). In both Me­ Archosauria to the (the Avesuchia sozoic birds and crocodyliforms, the teeth lack of Benton [1999]), but even so this statement is serrations, a constriction appears between crown incorrect. are present in “sphenosu­ and root, the root of the tooth is inflated, and in­ chian” crocodylomorphs and crocodylians (Re­ terdental plates do not develop (see Martin and ese 1915, Mook 1921, Walker 1972, Crush 1984, Stewart [1999] for an extensive overview). This Wu and Chatterjee 1993), and the interclavicle is situation contrasts with that in most theropods, present in carinate neornithine birds, where it is which have serrated, recurved teeth with exten­ incorporated into the sternum, contributing to sive interdental plates, as in archosaurs primi­ the carina (Parker 1891, Romanoff 1960, Martin tively. Martin et al. (1980), Martin (1983, 1985, 1991). Following the argument of Bellairs and 1991), and Martin and Stewart (1999) argued Jenkin (1960) that the hypocleidium of the fur­ that these data supported the crocodylomorph cula might be homologous with the interclavicle, hypothesis of bird origins first put forward by Martin et al. (1998a) argued that the exaggerated Walker (1972). Currie (1987) attempted to identify hypocleidium of Enantiornithes (in most taxa it similar features in the dentition of troodontids. is 75% the length of the furcular rami; Chiappe Elements of Martin et al.’s (1980) original argu­ and Walker 2002) was formed by the interclavicle. ment have been criticized by other researchers According to this argument, in enantiornithine who have sought to show that the dentition and birds the interclavicle was incorporated into the implantation style of the teeth in Archaeopteryx is furcula rather than into the sternum as a promi­ more similar to that of theropods than Martin et nent carina, as in carinate neornithines. al. (1980) realized (e.g., Elzanowski and Welln­ Norell and Makovicky (1999) argued that the hofer 1996, Elzanowski 2002). hypocleidium in birds could not be homologous In recent years, the sharp dichotomy between to the interclavicle because the interclavicle is ab­ the avian-crocodyliform dental morphology and sent in ornithodirans and birds are ornithodirans, mode of implantation on the one hand, and thero­ but this is a circular argument. They cited Russell pod dental morphology and mode of implantation and Joffe’s (1985) study of the early development on the other, has clearly become untenable. Numer­ of the furcula in Coturnix as supporting their view, ous maniraptorans have combinations of avian– but Russell and Joffe’s work only reiterates the crocodyliform dental characters (Weishampel et al. earlier conclusion (e.g., Heilmann 1926) that the 2004, and references therein), and some manirap­ interclavicle is not part of the furcula in neorni­ torans, like Byronosaurus, a troodontid, appear thines. Martin et al. (1998a) did not argue that the to share the entire suite of avian–crocodyliform hypocleidium and interclavicle were homologous dental characters (Makovicky et al. 2003; see also in all birds, but that they were homologous only Norell and Hwang 2004). Moreover, as is now within the enantiornithines. Thus, Russell and known from studies of mammals (e.g., Naylor Joffe’s study supports Norell and Makovicky’s and Adams 2001, and references therein), den­ (1999) argument that the interclavicle and hypo­ tal characters are considerably more plastic than cleidium are distinct elements only insofar as this once realized. Nonetheless, these characters are argument pertains to ornithurine birds. Padian potentially phylogenetically informative in dis­ (2001b) argued that, because the interclavicle was criminating among competing hypotheses about “lost or unknown” in most crown-group archo­ the origin of birds and, therefore, were included saurs, Martin et al. (1998a) were mistaken; but, in our analysis. as noted above, Padian was incorrect about the 3 Although interclavicle characters have not, to distribution of the interclavicle in crown-group our knowledge, been used in cladistic analyses archosaurs.

OM66_Text.indd 62 3/31/09 5:55:18 PM origin of birds 63

We coded the enantiornithine taxa in our anal­ and Gauthier 1999). In particular, the manus of ysis as having the interclavicle incorporated into maniraptorans is similar in overall morphology the furcula, and we coded carinate ornithurine to that of Archaeopteryx (e.g., Ostrom 1976a, Gau­ taxa as having the interclavicle incorporated into thier 1986, Wagner and Gauthier 1999, Paul 2002). the sternum. This similarity has lent support to the BMT hy­ 4 The homology of the avian furcula has been pothesis. In theropods, and in dinosaurs primi­ contentious (e.g., Bryant and Russell 1993, Hall tively, the pattern of digital reduction appears to 2001), in spite of the common assertion that the have been postaxial and asymmetrical with the furcula of birds is formed by medial fusion of reduction or loss of digits IV and V (Gauthier the clavicles of reptiles (e.g., Baumel and Witmer 1986, Wagner and Gauthier 1999). Both basal sau­ 1993). The furcula has long been considered a rischians and basal ornithischians show evidence character vital to an understanding of the origin of this pattern of digital reduction (Weishampel of birds. In his famous treatise, Heilmann (1926) et al. 2004, and references therein). Therefore, if argued that, because theropods lacked clavicles, the tridactyl manus of birds and neotetanurine they could not have given rise to birds, because theropods are homologous, then the manus of they could not have evolved a furcula. Structures birds should be composed of digits I, II, and III. identified as fused clavicles forming furculae The phalangeal formula of the manual digits of have now been reported in several theropod taxa, Archaeopteryx is 2-3-4, corresponding to the first including ceratosaurs (Tykoski and Rowe 2004), three digits of the pentadactyl manus of primi­ Allosaurus (Chure and Madsen 1996), tyrannosau­ tive archosaurs and supporting the contention roids (Makovicky and Currie 1998), oviraptoro­ that the manual digits of birds are I, II, III (Os­ saurs (Osmólska et al. 2004), troodontids (Xu and trom 1976a, Gauthier 1986, Wagner and Gauthier Norell 2004), and dromaeosaurs (Norell and Mak­ 1999, Paul 2002). ovicky 2004). The presence in theropod dinosaurs Unfortunately, a conflict between the paleon­ of structures possibly homologous with the avian tological and embryological evidence casts doubt furcula has been considered particularly signifi­ on the homology of the avian and theropod man­ cant (e.g., Padian 2001b, Paul 2002), but furcula- ual digits. Embryological evidence from Gallus like structures have also been identified in early and Struthio suggests that the manual digits of archosaurs like Longisquama (Sharov 1970; Jones birds are the second, third, and fourth of the pen­ et al. 2000, 2001) and prosauropod dinosaurs tadactyl manus of primitive archosaurs (Hinch­ (Yates and Vasconcelos 2005), and they also seem liffe and Hecht 1984, Hinchliffe 1985, Burke and to be present in at least one group of nonarcho­ Feduccia 1997, Feduccia 1999, Feduccia and saurian diapsids, the drepanosaurs (Harris and Nowicki 2002, Kundrát et al. 2002, Larsson and Downs 2002). Despite potential homology prob­ Wagner 2002, Feduccia et al. 2005; but see Welten lems, and because of the historical significance et al. 2005). Whereas in dinosaurs digits IV and of the furcula as a character in avian evolution, V were reduced or lost and the first, second, and we followed the consensus among ornithologists third digits of the pentadactyl manus of primitive and regarded the furcula as homologous with the archosaurs were retained, in birds the digital re­ reptilian clavicles. This decision may have to be duction pattern appears to be symmetric around revised in the future. digit III. To account for the discrepancy, Wagner and Appendix 3: Analysis of Characters Gauthier (1999) and Vargas and Fallon (2005a, b) in the Manus, Carpus, and Tarsus That have suggested that a homeotic frame shift must Were Excluded from the Primary have occurred during the evolution of neotetanu­ Analysis of Our New Matrix rine theropods, whereby the expression domains of genes such as the Hox d group were reposi­ Manus tioned in the limb bud (see also Wagner 2005). According to this hypothesis, embryonic con­ Among archosaurs, only birds and some neo­ densations II, III, and IV give rise in adult birds tetanurine theropods (the and Coe­ to digits corresponding to the first, second, and lurosauria, equivalent to of Holtz third digits of the pentadactyl manus of primitive et al. [2004]) have a manus with only three dig­ archosaurs, thereby restoring digital homology its (e.g., Ostrom 1976a, Gauthier 1986, Wagner between theropods and birds. Dahn and Fallon

OM66_Text.indd 63 3/31/09 5:55:19 PM 64 ORNITHOLOGICAL MONOGRAPHS NO. 66

(2000) and Drossopoulou et al. (2000) found that Feduccia et al. 2005). This ancestral character state a homeotic frame shift such as that proposed by is incompatible with the manual morphology of Wagner and Gauthier (1999) can be produced in basal saurischians and theropods, because these the laboratory by manipulation of the normal taxa are committed to a postaxial, asymmetrical development of the limb bud, and Vargas and pattern of digital reduction, with digits I, II, and Fallon (2005a, b) argued that molecular evidence III predominant (Wagner and Gauthier 1999, Fe­ supports the frame-shift hypothesis. Some mo­ duccia 2002, Feduccia et al. 2005). The frame-shift lecular studies have retrieved different results, hypothesis, even if accepted, does not explain however (Welten et al. 2005), and Galis et al. this discrepancy. (2005) and Feduccia et al. (2005) have shown that In an effort to resolve these uncertainties, Galis the molecular evidence advanced by Vargas and et al. (2003) suggested that neotetanurine thero­ Fallon (2005a, b) is inconclusive. Feduccia (2002), pods may have retained digits II, III, and IV in­ Galis et al. (2003, 2005), and Feduccia et al. (2005) stead of I, II, and III. This interpretation, if correct, argued that, even though it is theoretically pos­ would mitigate the discrepancy between the em­ sible, the frame-shift hypothesis is biologically bryological and paleontological data and would implausible, especially given constraints on auto­ render the frame-shift hypothesis unnecessary, pod development in and the absence of but Larsson and Wagner (2003) have criticized any clear adaptive advantage achieved through Galis et al.’s (2003) proposal. Given that a post­ dramatic alterations of normal autopod develop­ axial, asymmetrical pattern of digital reduction ment. Known examples of frame shifts in amniotes appears to be primitive within Dinosauria, the are on a smaller scale (Feduccia et al. 2005). Prum proposal of Galis et al. (2003) is difficult to sup­ (2002, 2003) has argued that if the digits of the port. avian hand are II, III, and IV, then explaining the Clearly, the present data concerning the identity phalangeal formula of Archaeopteryx will require of the manual digits of both birds and theropods postulating transformations during autopod de­ are ambiguous. Of course, the digits of birds and velopment that are similar to those postulated by theropods would be homologous if birds were the frame-shift hypothesis. If the digits of Archae- theropods, and the digits of birds and manirap­ opteryx were really II, III, and IV, a symmetrical torans would be homologous if maniraptorans reduction of one phalanx per digit would have were birds, but the data are not sufficient to estab­ had to occur to yield the observed phalangeal for­ lish these homologies at present, and any scoring mula of 2-3-4. Feduccia (1999, 2002, pers. comm.) of the manual digits for birds and theropods does and Feduccia et al. (2005) contended that this is a not acknowledge the uncertainties in the data. In simpler and more plausible transformation than addition, these uncertainties complicate compari­ a homeotic frame shift; it can occur via blockage sons of the avian and theropod manus with the of BMP4 signaling during limb bud development manus of archosaurs in general. (Zhou and Niswander 1996). Moreover, variation in phalangeal formulae is common in birds, even Carpus in the Mesozoic, and pedal phalangeal formulae even vary in some specimens of Archaeopteryx (Fe­ Evaluation of carpal homologies among birds duccia 2002, Feduccia et al. 2005). Feduccia (2002) and theropods is difficult. The highly derived car­ and Feduccia et al. (2005) further noted that the pus of neornithine birds is not easily compared frame-shift hypothesis fails to account for all of with the carpus of either Mesozoic birds or thero­ the embryological evidence. Demonstration of the pods. The considerable morphological variation presence of all five digits in the manus of birds, in the carpus of theropods is often poorly un­ with digits II, III, and IV prominent, indicates that derstood. Furthermore, homologies in the avian the manus of birds was primitively pentadactyl, and theropod carpus are contingent upon digital with digits I and V reduced, the common pattern identity, and the digital identities of theropods of digital reduction in amniotes. These embryo­ and birds remain uncertain. Even if this issue logical data for the normal development of birds is disregarded, seven more arise: (1) disappear­ indicate that the direct ancestral lineage of Aves ance of the avian ulnare and development of the was characterized by a pentadactyl manus with pisiform, (2) homology of the proximal carpals in digits II, III, and IV predominant (Feduccia 2002, Maniraptora, (3) distribution of proximal carpals Feduccia and Nowicki 2002, Kundrát et al. 2002, in nonmaniraptoran theropods, (4) homology of

OM66_Text.indd 64 3/31/09 5:55:19 PM origin of birds 65

the radiale within Aves and between birds and positions in some specimens, its identity is un­ theropods, (5) homology of the semilunate car­ certain (Osmólska et al. 2004). In Ingenia, a carpal pal within Aves, (6) homology of the semilunate associated with the topographically third meta­ carpal within Theropoda and between theropods carpal may be the corresponding distal carpal, and birds, and (7) identification of distal carpals but in Caudipteryx, a similar element is ventral without knowledge of digital identity. Compari­ to the ulna, and Zhou et al. (2000) labeled it an sons among birds, theropods, and other archo­ “ulnare.” Osmólska et al. (2004) were uncertain saurs for these characters are compromised by about its identity. With the exception of basal the ambiguity of the data. forms, the dromaeosaur carpus has only two ele­ Disappearance of the avian ulnare and develop- ments, a semilunate distal carpal and a proximal ment of the pisiform.—A V- or U-shaped proximal carpal. Ostrom (1969) mistook the semilunate bone in the neornithine carpus is identified as the carpal of Deinonychus for the radiale, but it is, in ulnare by Baumel and Witmer (1993). Hinchliffe fact, a distal element (Padian 2001b, Paul 2002); (1985) observed that the reptilian ulnare is pres­ he suggested that the other carpal in Deinonychus ent as a cartilaginous element in the embryonic was the ulnare, but it may be the radiale or even chicken carpus at 5.5 days of development but another carpal entirely; we presently have no that, by day 7.5, the true ulnare has stopped syn­ way of knowing (Norell and Makovicky 2004). thesizing matrix, undergone cell death, and dis­ The situation in troodontids is equally uncertain: appeared (Fig. 4). Subsequently, a V- or U-shaped the carpus is known only for Sinornithoides, which pisiform develops (Figs. 4 and 5). Hinchliffe and possesses two elements identified as a semilunate Hecht (1984) have detected this developmental carpal and a radiale (Russell and Dong 1993a). pattern in the embryos of anseriforms, charadrii­ Again, however, we have no way of knowing forms, galliforms, spheniscids, and Struthio. It whether this proximal element actually is the may be primitive for Neornithes. These conclu­ radiale. Given these data, the Oviraptorosauria, sions have been generally accepted in the orni­ Dromaeosauridae, and Troodontidae cannot cur­ thological literature, despite the persistence of rently be coded for these characters. the term “ulnare” on grounds of entrenched Distribution of proximal carpals in nonmanirap- usage (e.g., Baumel and Witmer 1993, Feduccia toran theropods.—Uncertainty about the distri­ 1999). In a recent study on the development of bution of proximal carpals in nonmaniraptoran the wing basipodium in Struthio, however, Kun­ theropods involves the ulnare, the intermedium, drát (2008) argued that the ulnare is replaced not and the pisiform. by the pisiform but by a proximal carpal element An ulnare is apparently present in the coelo­ of uncertain identity, which he identified as the physoid ceratosaur Syntarsus (= Coelophysis; see pseudoulnare. These data complicate assessment Paul 2002), along with a radiale and intermedium of the basal pattern for Aves and further compli­ (Tykoski and Rowe 2004), but neotetanurines may cate comparison of the avian carpus with that of primitively lack an ulnare, given that no ulnare is other archosaurs. preserved in the carpus of Allosaurus, although a No developmental data for the carpus of thero­ radiale and an intermedium are preserved (Chure pods are available to indicate whether a similar 2001). The situation in coelurosaurs is unclear. process occurred, and it is unclear whether any Hwang et al. (2004) identified a small discoidal car­ theropods possessed a pisiform, much less a pal between the ulna and radius in Huaxiagnathus pseudoulnare (see below). Assessing the homol­ as an ulnare, but this conclusion is unlikely. First, ogy of the avian “ulnare” and the theropod ulnare, the ulnare may be absent in neotetanurine out­ or of the avian pisiform and a theropod pisiform, groups to the Coelurosauria; second, the position or of an avian pseudoulnare and a theropod of this carpal between the ulna and radius is more pseudoulnare, if any of these elements were in fact congruent with its being an intermedium; and present in theropods, is therefore impossible. third, this position cannot be explained by post­ Homology of the proximal carpals in Maniraptora.— mortem distortion, because the carpal in question Confusion remains about the identities of the is preserved in the same position in both wrists of proximal carpal elements in Oviraptorosauria, the type specimen (CAGS-IG02-301; see fig. 8 of Dromaeosauridae, and Troodontidae. In ovirap­ Hwang et al. 2004). Similar considerations apply torosaurs, a carpal usually associated with the to the carpal identified as an ulnare inSinosaurop - radius may be the radiale, but because it shifts teryx by Currie and Chen (2001). Currie and Chen

OM66_Text.indd 65 3/31/09 5:55:19 PM 66 ORNITHOLOGICAL MONOGRAPHS NO. 66

(2001) also reported that a pisiform is present in to determine whether a “radiale,” if present in a Sinosauropteryx, but the element in question could theropod taxon, would be homologous with the equally well be a disarticulated manual phalanx, avian radiale. and a pisiform is apparently absent in all other Homology of the semilunate carpal within Aves.—A theropods. The situation in ornithomimosaurs is distal semilunate carpal is found in embryonic confusing. In their description of Sinornithomi- but not adult neornithine birds. During develop­ mus, Kobayashi and Lü (2003) stated that, among ment of the carpus, it fuses to the metacarpals the proximal carpals, an ulnare and an interme­ to form the carpal trochlea, a component of the dium are present, but their picture of the carpus carpometacarpus (Trochlea carpalis; Baumel and of IVPP V 11797-18 (Kobayashi and Lü 2003, fig. Witmer 1993). Hinchliffe documented the pres­ 16B) does not seem to corroborate this interpre­ ence of two elements in the distal row of the tation. Kobayashi and Lü’s “intermedium” is on carpus of the embryonic skeleton of the chicken the far medial side of the carpus, distal to the ra­ wing (see figs. 9–10 ofH inchliffe 1985; our Fig. 5). dius. It is not positioned between the radius and One element is associated with the middle meta­ ulnare, as the intermedium ought to be. Given the carpal (identified byH inchliffe as metacarpal III), tight articulation of the wrist in this specimen, the and Hinchliffe (1985:144) refers to it as “distal carpal in question is unlikely to have been dis­ carpal III.” The other element in the distal row placed. This carpal is probably the radiale, not the is of unknown identity. It is located just distal to intermedium. The structure that Kobayashi and the outer metacarpal (identified by Hinchliffe as Lü (2003) labeled “ulnare” is positioned distal to metacarpal IV); it was noted by Montagna (1945), the ulna but is also closely applied to the topo­ who labeled it “X.” Hinchliffe’s distal carpal III graphical third metacarpal. It could as easily be fuses to the outer metacarpal at its radial (me­ interpreted as a distal carpal. These ambiguities dial) edge, whereas at its ulnar (lateral) edge it complicate our understanding of the basal condi­ fuses with element “X.” Later, this semicircular tion for Ornithomimosauria and, by extension, of structure fuses to the inner metacarpal and then, the basal condition for Maniraptoriformes. finally, to the middle metacarpal, forming the These data do not make clear which proximal crescentic carpal trochlea in the adult bird (Fig. elements were retained in nonmaniraptoran thero­ 5). Kundrát (2008), on the basis of a study of the pod evolution. Some elements, like the ulnare and development of the wing basipodium in Struthio, intermedium, may have been lost and reacquired argued, against Hinchliffe (1985), that the semi­ repeatedly. Determination of whether they were lunate carpal is formed from the fusion of distal is complicated by uncertainties in interpreting the carpals II–IV, and he suggested that this pattern available fossils. A pisiform does not appear to is primitive for Aves. No means is currently avail­ be unambiguously present in nonmaniraptoran able to determine whether he is correct. or maniraptoran theropods. Until further data In Archaeopteryx and other basal birds, a prom­ clarify trends in the evolution of proximal carpals inent semilunate carpal persists in the adult within the Theropoda, coding theropods for these carpus. It remains free of the metacarpus in Ar- characters cannot be justified. chaeopteryx, but in some basal birds, like Confuciu- Homology of the radiale within Aves and between sornis, it may be partially fused to the metacarpus. birds and theropods.—On the basis of a study of If Aves is monophyletic, then the process by the development of the wing basipodium in Stru- which the semilunate develops in neornithine thio, Kundrát (2008) has argued that the radiale birds may be primitive to all birds, but Zhou and of neornithines is a composite element formed Martin (1999:291, fig. 3) have identified a small by the fusion of the radiale sensu stricto and the free carpal (possibly a distal carpal) in Archaeop- intermedium. Assessment of whether the radiale teryx as the element “X” identified by Hinchliffe of Mesozoic birds is a composite element is im­ (1985), stating that only in modern birds does ele­ possible without developmental data. Therefore, ment “X” fuse to the semilunate carpal. If so, the the homology of the radiale within Aves cannot semilunate of Archaeopteryx formed differently be assessed. Given the uncertainty with respect from the semilunate of modern birds, and the to the distribution of the intermedium and radi­ two may not be homologous. If Kundrát’s (2008) ale sensu stricto within Theropoda (inclusive of analysis of the development of the avian semi­ maniraptorans; see above), combined with the ab­ lunate carpal is correct, however, the semilunate sence of developmental data, it is also impossible carpals of Archaeopteryx and modern birds may

OM66_Text.indd 66 3/31/09 5:55:20 PM origin of birds 67

be homologous. Without further data on the ho­ be homologous with the semilunate carpals of mologies of the elements that form the carpal tro­ maniraptorans (Chure 2001). Xu et al. (2006) con­ chlea of neornithines, and without data clarifying curred with Chure (2001), but Holtz (2001) sup­ homologies in the carpus of Mesozoic birds, the ported the homology of the semilunate carpals in homology of the semilunate carpal within Aves nonmaniraptoran and maniraptoran theropods is, at present, uncertain. and considered the semilunate diagnostic of a Homology of the semilunate carpal within the large clade of theropods, perhaps the Tetanurae. Theropoda and between theropods and birds.—Many Establishment of the homology of the semi­ researchers have identified structures in the lunate carpal within the Theropoda is further theropod carpus as semilunate carpals formed complicated by uncertainties concerning the from the fusion of distal carpals I and II, and basal conditions for Neotetanurae, Coelurosau­ they have considered them homologous with the ria, Maniraptoriformes, and Maniraptora. The semilunate carpals of birds. Semilunate carpals putative semilunate carpal of Allosaurus may be have been identified in the basal tetanurines Xu- a single distal carpal (Chure 2001); if so, it would anhanosaurus (Holtz et al. 2004); the spinosauroid not be a semilunate carpal by the criteria of either (Sereno et al. 1994); possibly the neo­ Padian and Chiappe (1998) or Chure (2001), so a tetanurine Allosaurus Chure (2001); the basal co­ semilunate carpal may be primitively absent in leurosaurian compsognathids Huaxiagnthus and the Neotetanurae. If compsognathids were basal Sinosauropteryx (Currie and Chen 2001, Hwang et maniraptorans (Hwang et al. 2004) rather than al. 2004); the basal tyrannosauroid Guanlong (Xu basal coelurosaurs (Forster et al. 1998, Sereno 1999, et al. 2006); possibly the derived ornithomimosaur Holtz et al. 2004), a semilunate carpal might be Struthiomimus (Nicholls and Russell 1985, Bars­ primitively absent in coelurosaurs, although Xu et bold and Osmólska 1990); the therizinosauroids al. (2006) reported the presence of a putative semi­ Falcarius, Alxasaurus, and (Russell lunate carpal in the basal tyrannosauroid - and Dong 1993b, Xu et al. 1999b, Kirkland et al. long. Even if a semilunate carpal were primitively 2005b, Zanno 2006); and the Oviraptorosauria, present in coelurosaurs, it does not appear to be Dromaeosauridae, and Troodontidae (e.g., Mak­ primitively present in Maniraptoriformes, given ovicky and Norell 2004, Norell and Makovicky that no basal ornithomimosaurs possess one (Bars­ 2004, Osmólska et al. 2004). The carpus of the bold and Perle 1984, Barsbold and Osmólska 1990). alvarezsaurids is too derived to be interpretable Although the basal therizinosauroid Falcarius was because the carpals are fused into a carpometa­ described as having a semilunate carpal, in some carpus (Chiappe et al. 2002). specimens no semilunate carpal is present (Zanno Establishment of the homology of the semi­ 2006), which suggests that therizinosauroids may lunate carpal within the Theropoda is difficult. primitively lack a semilunate carpal. If they do, the Only in the maniraptoran clades Oviraptoro­ basal condition for Maniraptora is unclear. These sauria, Dromaeosauridae, and Troodontidae is ambiguities clearly indicate that the homologies a prominent semilunate distal carpal with tro­ of the semilunate carpal within Theropoda cannot chlear surfaces present. In nonmaniraptoran presently be resolved. theropods, the elements identified as semilunate Homology of the semilunate carpals of thero­ carpals are small and only vaguely semilunate, pods and birds is also difficult to establish. Al­ and they lack trochlear surfaces. Padian and though Chure (2001) argued that a trochlear Chiappe (1998) regarded fusion of distal carpals I surface must be present for a carpal to count as and II sufficient to indicate that a semilunate car­ a semilunate, he did not dispute the widespread pal homologous to that of birds is present. Note, assumption that semilunate carpals of thero­ however, that although assumptions of fusion pods are composed of the fused distal carpals I in carpalia are common, they are often inferred and II (e.g., Padian and Chiappe 1998), but if the without any evidence; genuine loss may be as semilunate carpals of theropod taxa, including common as or more common than fusion of sepa­ maniraptorans, are composed of fused distal car­ rate elements (Romer 1956). Contrary to Padian pals I and II, then the theropod semilunate carpal and Chiappe (1998), Chure (2001) argued that a cannot be homologous with the semilunate car­ trochlear surface must be present for a carpal to pal of neornithines, which is composed of distal be considered a semilunate. If so, the “semilunate carpal III and an unidentified lateral element X carpal” of nonmaniraptoran theropods cannot (Fig. 5; Hinchliffe 1985). Even if Kundrát (2008)

OM66_Text.indd 67 3/31/09 5:55:20 PM 68 ORNITHOLOGICAL MONOGRAPHS NO. 66

were correct in arguing that the semilunate car­ this structure the “pretibial.” Ostrom (e.g., 1976a, pal of neornithines is composed of fused distal 1985) argued that this structure is homologous carpals II–IV, the semilunate carpal of theropods with a similar structure in the tarsus of theropods would still not be homologous with the avian (see also Paul 2002), but according to Martin et semilunate carpal if, as Padian and Chiappe al. (1980:88) “differences in placement and [the (1998) and others have asserted, the theropod pretibial’s] late appearance during development semilunate is composed of fused distal carpals I suggest that it is a uniquely derived character and II. The similarity of the semilunate carpals of for birds and is properly termed a pretibial bone, oviraptorosaurs, dromaeosaurs, troodontids, and rather than an astragalar process.” In contrast to Mesozoic birds is not, by itself, sufficient to imply the situation in neornithine and Mesozoic birds, homology, and the homologies of the semilunate the ascending process of theropods is usually a carpals in these taxa cannot presently be indepen­ broad sheet of bone, continuous and exclusively dently assessed. Therefore, even if the uncertain associated with the astragalus (compare Fig. 6A digital identities of theropods and birds are disre­ and B). garded, the homology of the theropod and avian Seeking to resolve this controversy, McGowan semilunate carpals clearly cannot be established (1984) claimed that the “pretibial” in ratites is ac­ at present. tually a continuation of the astragalus and that it Identification of distal carpals without knowledge is not a separate ossification subsequently fusing of digital identity.—Distal carpals are identified with the astragalus. He claimed that the “pretib­ by reference to the digits with which they are ial” ossification was a part of the cartilaginous associated, so which distal carpals are present precursor of the calcaneum, which he labeled the in birds or theropods cannot be determined be­ “calcaneal spur.” This “pretibial” ossified and fore the digital identities of birds and theropods fused with the astragalus in ratites and with the have been ascertained. Clearly, neither birds nor calcaneum in carinates. McGowan (1985) later theropods can be scored for these characters until modified his position and argued that, in both the digital identities of birds and theropods have and carinate birds, the “pretibial” was a de­ been resolved. rivative of the astragalus, and that the chief dif­ ference was the point of fusion, a shift laterally Tarsus: Ascending Process toward the calcaneum being a derived condition of the Astragalus in birds (see also Paul 2002). In either version, McGowan’s conclusions would, if correct, obvi­ In birds and theropods, a sheet of bone that ate the apparent discrepancy between the mor­ braces the anterior face of the tibia is usually phology of the avian and theropod ascending called the “ascending process of the astragalus” processes. or simply the “ascending process.” It is less evi­ Martin and Stewart (1985) challenged McGow­ dent in adult neornithines than in juvenile (or em­ an’s findings, observing that his use of late-stage bryonic) neornithines and Mesozoic birds. This embryos obscured the actual relationships of the sheet of bone is particularly conspicuous in basal bones in question. Whereas the youngest em­ birds, including Archaeopteryx. This common fea­ bryos McGowan (1984) examined were 12 days ture has consistently been regarded as one of the old, Martin and Stewart (1985) examined seven- most striking homologies shared by birds and day-old embryos of with completely theropods (e.g., Paul 2002), but comparative ana­ distinct and still cartilaginous astragali, calcanea, tomical research reveals that establishing the ho­ and tibiae. At eight days, a triangular pretibial mologies of the ascending processes of theropods appears, as described by Morse (1872), and birds is difficult. and by nine days this pretibial cartilage fuses In neornithines a triangular, late-developing with the calcaneal and astragalar cartilages. cartilage appears, after fusion of the proximal Martin and Stewart (1985:160) concluded that tarsals, on the lateral face of the tibia, dorsal “ratites and carinates have an ascending process to the calcaneum (Martin et al. 1980, and refer­ (pretibial bone) separate from both calcaneum ences therein). Subsequently, this cartilage fuses and astragalus, but that the position of this pro­ with the calcaneum, with which it is primarily cess may vary.” They reaffirmed the argument of associated in both Mesozoic and modern birds Martin et al. (1980) that the pretibial ossification (Martin et al. 1980; Fig. 6). Morse (1872) called is an avian neomorph. Feduccia (pers. comm.)

OM66_Text.indd 68 3/31/09 5:55:20 PM origin of birds 69

has confirmed these results with embryos of Archaeopteryx “the astragalus forms a broad as­ Struthio and suggested that the term “descend­ cending process identical to that of theropod ing process” be used to describe the structure in dinosaurs” (Mayr 2005:1485). This statement birds more accurately. Holmgren (1955) also re­ contrasts with the account of Martin et al. (1980), ported that, in the embryonic tarsus of Struthio, who stated that the pretibial of Archaeopteryx was the pretibial was exclusively associated with the oriented laterally, not medially, and that it was calcaneum. therefore primarily associated with the calca­ Welles (1984) claimed that the ascending pro­ neum and not the astragalus. The figures of the cess in theropods was actually a separate ossifica­ tarsus in the “Thermopolis specimen” provided tion, as in birds, an interpretation based on the by Mayr et al. (2005, fig. 3C; 2007, fig. 12) do not crus of a juvenile specimen of Dilophosaurus, a unambiguously support their interpretation. coelophysoid ceratosaur (Welles 1984:140, fig. 35). Close examination of figure 12b from Mayr et Paul (2002) concurred with Welles’s assessment al. (2007), showing the anterior face of the right and suggested that the only difference between tarsus, reveals a spear-shaped extension of bone the ascending process in birds and the ascending contiguous with the sheet of bone identified as process in theropods was its more or less medial the ascending process. This spear-shaped exten­ position (more medial in “nonavian theropods” sion is clearly part of the “ascending process” and less medial in birds). He argued that the de­ and is positioned laterally, nearer the calcaneum velopment of a supratendinal bridge and exten­ (Fig. 6C). As noted above, the pretibial in birds sor canal for the tendon of M. Extensor digitorum develops as a separate center of ossification, longus had shifted the pretibial laterally in mod­ which then expands distally and fuses to the cal­ ern birds, but in at least some Archaeopteryx speci­ caneum, sometimes, as in ratites or falconiforms, mens, and in other Mesozoic birds, the pretibial is developing subsequent contact with the medial already laterally positioned, despite the absence astragalus. The shape of the ascending process of a supratendinal bridge (Martin et al. 1980). and its spear-shaped extension in the tarsus of In the absence of embryological data for thero­ the “Thermopolis specimen” is consistent with pods, it is not possible to determine whether the this developmental process: initial development structure in theropods is actually a continuation of the pretibial, subsequent ventral growth, and of the astragalus or, as per Welles (1984) and Paul contact with the astragalus through medial ex­ (2002), a separate, pretibial ossification homolo­ pansion. The situation is different from that seen gous with the avian pretibial (“descending pro­ in a typical theropod tarsus, such as that of Al- cess”). Moreover, uncertainty remains about the bertosaurus, in which the ascending process is an distribution of this character in other archosaurs. uninterrupted sheet of bone that dominates the An ascending process is present in dinosauro­ entire lower anterior surface of the tibia (com­ morphs such as Marasuchus (= ; Welles pare Fig. 6A and B with C). Furthermore, the and Long 1974; fig. 10.14d of Paul 2002) and in evident line separating the ascending process the ornithischian (Galton 1974). An and proximal tarsals in the “Thermopolis speci­ ascending process is also present in sauropods men” has the appearance of a suture, in agree­ (Upchurch et al. 2004). Were we to accept the ment with the assessment offered above. On the small bump on the astragalus of Marasuchus as an basis of the evidence available, the exception to incipient, but nonetheless discernible, ascending Martin et al.’s (1980) description of the tarsus of process (Paul 2002, fig. 10.14d; Sereno and Arcucci Archaeopteryx is that, in at least one specimen of 1994:64, fig. 10E), we would have as much cause Archaeopteryx, the pretibial has developed some to identify a similar spur of bone on the astrag­ medial contact with the astragalus. alus of crocodylians as an ascending process (e.g., Clearly, the homologies of these structures in fig. 20 of Mook 1921, fig. 10.14C of Paul 2002).W e birds and theropods remain uncertain. do not mean to say that crocodylians possess an ascending process in the sense that theropods do, Appendix 4: Data Matrix but this example illustrates that one should use caution in talking about small bumps on little The matrix includes 242 characters, of which ankle bones. the first 221 were used in the primary analysis. In a recent development, Mayr et al. (2005, 2007) The last 21 characters in each entry were turned reported that in the “Thermopolis specimen” of on only for the alternative analysis.

OM66_Text.indd 69 3/31/09 5:55:21 PM 70 ORNITHOLOGICAL MONOGRAPHS NO. 66

Alligator 00000 00000 00000 0000? 001?? 10000 00301 11110 11101 12122 01112 01??1 10101 11210 41001 11110 0—1 00000 00020 00201 00001 01011 00000 01112 00100 00001 00111 12121 21102 21000 21301 11101 02-01 10001 0-120 00000 01111 00011 11101 11110 31100 00000 10000 01021 0???? ????? ????? ????? ?? 00202 2001- 21111 10000 00000 00000 01110 10100 00001 00000 00100 00000 0—- -01– 20000 02000 Baptornis 00000 10001 00002 00010 0-001 10010 100-0 11000 ????0 ????? ????? ????? ????? ????? ??112 ?0??? 1???? 10010 10001 11100 00010 0?000 00011 00000 00000 ????? ????? ????? ????? ????? ????? ????? ????? ????? 0? ????? ????? ?0??? ???11 10?11 10111 14001 00104 Allosaurus 120-0 12111 2?1?1 1???? 0???1 02101 10010 00020 00111 10020 0-211 2?100 000-1 ?1112 -0101 21200 ????0 00000 00011 00100 00111 00000 00000 00000 10030 21030 2???? ????? ????? ????? ?? 00000 10000 01010 01110 01101 00001 00000 000-0 00000 00000 00000 00000 00100 10000 00001 00100 Byronosaurus 00100 01012 00100 00000 0—- -1200 21000 00000 ????? 00101 01111 ?1??? ?100? ????? 1???? ????? 00001 10000 00001 01001 10002 11000 10100 01100 ????? ????? ??120 ????? 111?1 100?1 11101 11100 20010 00000 10000 01020 0???? ????? ????? ????? ?? 00202 2111- 1111? 10000 00101 101?? ???10 ??1?? Alxasaurus ?1101 01?1? ??0?? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? 1???? ????? ????? ????? ????? ????? ????? ????? ????? ????? ?0??? ????? 0???1 ????? ????? ????? ????? ?? ????? ????? ????? ????? ????? ????? ????? ?0??? 01000 0??1? 0021? 10??? ????? ???0? ??100 01100 Caudipteryx 01012 1?110 01?00 0—- -11– 21000 00??0 00?01 00000 00000 02121 21002 2???? ?0211 11100 31??0 11000 0001? 11011 00012 -1011 01000 01100 0???1 0??0? ????0 010?0 ????? ????? ????? ????? ?? 01000 10000 0???? ????? ????? ????? ????? ????0 1120- —– 21–1 11211 00000 00010 ??0?? ??1?0 ?1??? Apsaravis ????2 ?20-0 0?0?? 1000? ????? 01000 10000 00001 ????? ????? ????? ????? ????? ????? ????? ????? ????? 0002? ???01 0102? ??012 01010 ?0201 11101 3?1?0 ????? ????? ????? ????? ????? ????? ????? ?1??? —-? 00000 10000 01020 1???? ????? ????? ????? ?? ???-? 1200? 0?1?? ????? ????? ??1?0 1400? 0???4 Ceratosaurus 120-0 12?0? 22??1 1???? 411?1 22101 00112 00112 00112 1002? ??212 20100 000-? 01102 -0101 ??200 ????0 00100 00011 11100 00111 00000 00000 00002 10020 11030 ????? ????? ????? ????? ?? 00001 00001 0?1?? ?1100 01101 010?1 00?00 000-0 00000 00000 00000 00000 00100 10000 00101 01100 Archaeopteryx 06110 11013 10100 01000 0—- -01– 21000 00000 12010 10100 01011 10012 -1001 ????? 11100 01111 00001 00000 00011 01002 0-002 11000 ?0120 01102 0??00 10000 11120 11111 11??? 000?1 11101 21100 -0010 00200 10000 11020 0???? ????? ????? ????? ?? 00202 2000- 21111 10000 0001- 00010 0?00? 101?0 Citipati ?10?0 02?12 02110 00000 0—- -1210 41011 21010 00000 00011 00?02 10121 11011 21000 210-0 00001 ????0 01010 12011 11002 -2011 01000 11100 01012 31100 00000 10020 01021 1???? ????? ????? ????? ?? 00000 10001 01121 12111 11101 010?1 00?01 11111 11— —– —– 11211 00001 00000 1?010 10110 15101 Avimimus 02123 11100 00011 10111 11211 41000 11000 00001 ????? ?1?1? ?201? ??012 -?0?1 012-2 13000 10110 00002 00101 00010 0-212 01010 10201 01101 30100 02-11 10011 1???? ????? 11111 010?? ?1??1 ????1 00000 10000 01020 0???? ????? ????? ????? ?? ?1–? —– ?-?-? -?2?? ?0?00 00011 10310 10101 11101 0101? ????? ????? ????? ????? ????? ????? ????? Coelurus 00012 0010? ?0021 0-212 11000 10??? ?1110 21001 ????? ????? ????? ????? ????? ????? ????? ????? ????? 20200 10020 11220 ????? ????? ????? ????? ?? ????? ????? ????? ????? ????? ????? ????? ?0??? ?000? ????? 0???? 01??? ????? ???0? ??110 11111 00?1? Bambiraptor ????1 0???? ????? ????? ???0? ????? ????? 00001 ????0 00000 01011 11002 -1001 01000 01000 10001 00112 0???? ????? ??000 2???? ?1110 2000? 0?000 00000 11110 0?1?? ?111? 1??0? 0???1 11?01 11100 ????? ????? ????? ????? ????? ????? ??

OM66_Text.indd 70 3/31/09 5:55:21 PM origin of birds 71

Compsognathus 01100 16110 01111 0?100 0???0 —?? -11– 20000 00000 00001 00000 00011 00002 0-002 11000 00120 0—0 00100 00011 00000 ?0001 01?00 ????? ????? 01100 10000 00000 10000 01020 0???? ????? ????? ????0 10000 0?1?? ????? 1???? 000?? ????? ????0 ????? ?? 00101 20100 ?000? 00000 0001- 00010 000?0 ?01?? ?2?0? ?001? ??100 00000 0—- -11– 21000 0???? Dromaeosaurus ????? ????? ???11 0???? ????? 11000 20100 0???? ????? ??00? 00011 ??000 1?0?1 ???0? 01000 10001 ????0 2?00? ??010 00020 0???? ????? ????? ????? ?? 00000 1111? 011?0 01110 01101 001?1 00001 10000 Conchoraptor 00000 00000 00000 10000 00110 10100 003?? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? 01010 12011 11002 -2011 01000 11100 01010 ????? ????? ????? ????? ????? ????? ????? ????? ????? 00000 10001 01121 12111 ????? ?10?? ????? ????1 ????? ????1 ????? ????? ????? ????? ?? 11— —– —– 11211 00001 00010 1?31? ??1?? ?1?0? 02124 1???? ??0?1 ????1 ??211 41000 11000 00001 Effigia 00002 00??? ?1011 0-212 01010 10??1 ?1101 301?? ????? 10100 11010 00010 00001 01110 1??00 0???1 ?000? ????? 0102? 0???? ????? ????? ????? ?? ?0000 10000 1??1? ????? ?1??? ????? ????? ????1 Confuciusornis 11— —– —– 11000 00000 00010 0?10? ??1?0 1???? 00?11 00101 00000 0—- -11– 00000 01000 00001 12000 11100 11010 00012 -2001 01101 11100 0?111 0002? ???11 ?001? 10001 11000 200-? 0110? ????0 02-01 10001 1???? ????? ????? ????? ????? ????1 ?0001 111?0 01010 0?000 00??? 00?00 00??? ?? 11— —– —– 11101 00000 00111 1?0?? ??1?? ?3001 12124 120-0 11001 20111 10210 01010 12100 00?00 Enaliornis 10002 00101 00120 0-111 21000 010-1 01112 -0101 ????? ????? ????? ????? ????? ????? ????? ??1?? ?2?11 00200 10020 11020 2???? ????? ????? ????? ?? 10001 1???? ????? 11111 00011 11111 2111? ????? Deinonychus ????? ????? ????? ????? ????? ???0? ??1?1 14001 00004 1???? ????? ????? ????? ????? ????? ????? ????? 00000 00011 10100 11001 01000 0000? ??001 ????? ????? ??0?? ??11? 2???? ????? ?1102 -010? ?000? ????? 011?0 01111 ????? ?0??? ????? ????0 ?1200 10030 21030 1???? ????? ????? ????? ?? 00000 00000 00000 1000? 00100 101?0 003?1 10110 11101 1212? ??112 010?? ????? 101– 21000 11010 Eoenantiornis 00101 00001 00111 00021 21002 21000 10201 0???1 12100 10100 11010 00012 -???1 1—- ????1 0???1 31100 00000 10000 01021 0???? ????? ????? ????? ?? ?2-1? 1000? 1???? ????? ????? ????? ????? ????0 Dibrothosuchus 00202 2002- 21111 11000 0001- 00011 ??0?? ????? ????? ????4 12??0 1?100 20011 13222 41011 22101 ????1 00000 00010 00001 00011 01000 00000 01112 01112 10122 00101 0?12? ??212 21100 210– -1112 02-00 00001 00120 00000 01111 00011 11101 11110 -2101 00200 10020 11030 2???? ????? ????? ????? 00201 10000 ?011? 00000 00000 00000 01310 00100 ?? 0100? 00000 0?0-0 00??? 0—- -?1– 20000 12000 00101 10002 00101 0?010 ??000 1???? ????? ????? Eoraptor ????? ????? ????? ????? ??000 00111 00000 000?0 0? ????0 00000 00010 00000 00001 00000 00000 00000 00000 00000 0???? ????? ????? ????? ????? ????0 Dilong 0?202 ????0 ?11?? ????0 00000 000?0 ??000 00100 0–?? 00000 01011 01000 10011 01010 00000 01??1 0?10? 00?01 00101 00000 0—- -11– 10000 00000 00000 00001 1???? ????? ???0? 00??? ????? ??1?0 00001 00000 00001 00001 0-201 11000 0010? 00000 00010 000?0 00000 0?000 011?? 110?0 ?1101 ??100 00000 00000 10000 01010 0???? ????? ????? ????? ?? 12??? ??01? ??100 ?0000 0—- -11– 21000 01000 Erlikosaurus 00001 00000 0???? ?0021 1?0?? 11000 2020? 0110? 2???0 0000? ??000 01020 0???? ????? ????? ????? ?? ????? 00100 01010 00000 00001 01000 01000 00001 10000 10001 01120 0?1?? 01111 0001? ???01 10101 Dilophosaurus 00–0 01000 0-110 00210 10000 00011 110?? ????? ????? 10000 00010 00100 0000? 00000 00000 00000 ????? ????? ????? ????? ????? ????? ????? ????? 0100? ????? ??1?? ????? 01101 ?00?1 00?0? 000-0 ????? 0?000 ????? ????? ????? ????? ????? ????? 00000 00000 ?000? 00000 00100 10000 00001 ????? ????? ???0? ?1?20 0???? ????? ????? ????? ??

OM66_Text.indd 71 3/31/09 5:55:21 PM 72 ORNITHOLOGICAL MONOGRAPHS NO. 66

Erpetosuchus 00010 00000 00000 0???0 01100 000?0 0?10? 001?0 11100 0001? 0?1?0 00?0? 0—- -11– 21000 00000 ????1 00000 00010 00200 00001 00011 00000 00??2 00?01 0000? ???01 01011 20002 11000 10000 01100 02-01 00001 00120 0000? ?0??? ????? ????? ????0 210?0 00000 100?0 01020 1???? ????? ????? ????? ?? 00202 2000- ?000? ????0 00000 00010 ??100 001?0 0100? 0000? ????? ????? 0—- -00– 21000 01000 Harpymimus 00-0? 10000 00??2 0???? ????? ????? ????? ????? ????? 00001 00011 00000 01001 00??? ????0 0???1 ????? ????? ????? ????? ??000 00??? 00000 00??? ?? ?0?00 00000 0???? ????? ????? ????? ????? ????1 Erythrosuchus 10— ?000? ?–?? 00010 00000 00011 0??0? ??1?0 ??10? 02?23 0?101 01?00 0—- -11– 11000 01000 ????1 00000 00010 00100 00000 01000 00000 00000 00?01 00020 00011 0?002 ??102 1?0?? ????? ?110? 00000 00000 0?1?? ?0000 01000 00000 00000 00000 ????0 ??000 10000 01030 3???? ????? ????? ????? ?? 00000 00000 00000 00000 00000 00000 00100 00000 05000 01010 01100 000?? ????? ????? 10000 Herrerasaurus 00000 0000? 00000 00000 00010 0-000 00000 000-0 ????? 00000 00010 00000 00001 00000 00000 00000 00003 -0000 00000 00000 00000 0?000 00??? 00000 01000 00000 0?100 011?0 01101 000?? 0??00 ????0 00??? ?0 00??1 10000 00000 00000 00100 101?0 ??300 00100 Euparkeria 00100 00000 00101 0?000 0—- -11– 21000 00000 00001 00001 00011 00010 0-102 11000 10100 01100 ????1 00000 00010 00000 00000 00000 00000 00000 00000 00000 10000 01010 0???? ????? ????? ????? 00000 10000 0?0?? ?0000 01000 00000 00000 00000 ?? 00100 10000 00000 00000 00000 00000 00300 00000 0100? 0???0 0?0-0 0?00? 0—- -00– 11000 00000 00000 Hesperosuchus 00010 00001 00020 0-000 00000 000-0 01003 -???0 ????1 00000 00010 00201 00001 00011 00000 00102 ?0000 00000 00000 0?000 00??? 00000 00??? ?0 01-00 00000 0???? ?0??? 11?1? ?00?? ????? ????0 Falcarius 00000 00000 0111? 0100? 000?? 0?0?0 ??000 00100 01001 0000? ??0-0 ?0000 0—- -0??? 10000 01000 ????? ??0?? 0???? ???0? ????1 01??? ????? ?1??? ????0 00001 00002 00111 0???? ????? ????? ????? ?1100 ?000? ????? ????? 11100 ?00?1 00?01 101?? 00??0 0000? 0000? ????0 01010 ????? ???11 00000 00000 0? 01000 1?11? 0100? 00??? ????? ????? ??10? 111?? ???22 0?101 0??00 0—- -1?– 21000 01000 00001 Heyuannia 00001 00111 02111 21002 11000 10301 11100 2???0 ????? ????? ????? ????? ????? 01??? ??100 01??? 00000 100?0 00020 0???? ????? ????? ????? ?? ????? ????? ????? ????? ????? ????? ????? ????? ?1??? Gallimimus —-? ???-? 1???1 00001 00011 ??11? ??1?? 1???? ???14 1?110 ?2011 1?1?1 ??211 41110 01010 00000 ????? 00001 00011 00010 01001 00111 10000 01001 0000? ??211 ?11?? ??112 01010 ?020? 0110? ????0 00000 00001 01120 0001? 01111 100?1 00?01 10101 0???? ??0?0 01020 1???? ????? ????? ????? ?? 11— —?- —– 11000 0001- -0010 ??101 10110 15100 02012 11001 01?00 0—- -11– 21000 01000 00001 Hongshanornis 00020 00011 01002 10202 11000 10100 01110 20010 12100 1?100 01??? ??012 -???1 ????? 1???? ????? 00000 10000 01130 3???? ????? ????? ????? ?? ????? ????? 1???? ????? ????? ????? ????? ????1 Gansus 11— —– —– 1???0 0?01- ????? ??0?? ????? ?1??? ????? 1???? 1?010 20011 11231 41011 22101 10?12 12100 ????? ????? ????? ????? ????? ????? ????? ????? 0111? ???0? ??12? ????? 21100 00??? ????? ????1 ????? ????? ????? ????? ????? ????? ????? ????? ????? 0?200 100?0 21030 2???? ????? ????? ????? ?? ????? ????? ????? ????? ???1? ??1?0 1?00? 02?24 120-0 1?100 22011 12230 41111 22101 10012 01112 Huaxiagnathus 11112 10120 0-212 21100 010-? 0???2 -0101 01200 ????? 0?000 00011 00??? ??00? 0?0?? 0???? ????? 10031 21030 2???? ????? ????? ????? ?? ????0 ????? 0???? ????? ????? ????? ????? ????0 00??1 10100 00000 ????0 0001- ??0?? ???00 001?0 0??0? Guanlong ????? ??100 00000 0—- -1200 21000 00000 00001 ????? 00000 01011 11000 10011 01000 000?? ????1 0002? ????1 00001 20002 11000 2010? 01100 ????0 ?000? 10001 0???? ????? 01?0? ????? ????? ????0 00000 10000 01020 0???? ????? ????? ????? ??

OM66_Text.indd 72 3/31/09 5:55:22 PM origin of birds 73

Iberomesornis 00?0? 0002? ???00 0???? ????? ????? ????? ????? ????? ????? ????? ????? ????? ??0?? 00000 00??? ?? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? Marasuchus ????? ????? ????? ????? ????? ??1?? -20?? ?0?02 ????? ????? 0???? ????? ????? ????? ????? ????? ????? 020-0 12101 2?1?? ?3222 ?11?1 12101 01?1? 0011? ????? ????? ????? ??1?1 ?0?01 00?0? ????? 0???? ??11? ?012? ????? 2???? ??0-? ????? ????? ?0000 ????? ????? ????? ????? ????? ???00 011?0 0?00? 10000 01030 2???? ????? ????? ????? ?? 00?00 00100 0???? 0—- -11– 00000 00000 0000? Ichthyornis 00000 ???00 00010 0-101 11000 000-0 01000 000?0 00000 10000 01010 0???? ????? ????? ????? ?? ????? 11100 1???? ??012 -???1 1—- ??112 01111 12-11 10001 1???? ????? ?11?? ???11 11?01 211?1 00–2 Microraptor 2001- 2-111 10–0 0011- 00011 11010 10111 13001 12010 10100 01??? ??01? ????? 0???? ????? ????? 12?24 120-1 1?101 21111 12230 41111 22101 10012 ????? ????? 1???? ????? ????? ????? ????? ????0 00201 01102 11112 10??? 0-21? 21100 010-? 01112 -010? 10000 ?111? 1000? 10??? ??1?? ???11 ??1?1 1??01 ?1210 10031 21030 2???? ????? ????? ????? ?? 10?03 11112 01011 20111 11210 41011 11110 00100 Incisivosaurus 0000? ?011? ?1111 2111? 21001 21301 11101 31100 00200 10000 01121 2???? ????? ????? ????? ?? ????? 01010 01011 11002 -2011 01000 00000 01??0 ?0000 10001 01120 0111? 11?0? ?10?? ????? ????0 Microvenator 00202 ?010- ?010? 02210 00000 00001 ??1?? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ??211 00??? ????? ???00 101?0 11101 0101? ????? ????? ????? ????? ????? ????? ?? ??0-0 0???? ????? ????? ????? ?0000 00001 00000 Ingenia 00011 000?1 20012 1?010 1???? ??100 2010? 0?000 100?? 0???? ????? ????? ????? ????? ?? ????? 01010 12011 ?1002 -2011 01000 10110 01011 00001 10001 0???? ????? ????? ????? ????? ????1 Mononykus 11— —– —– 11211 00001 00011 ??1?? ????? ????? ????? ????? 0???0 ????? ????? ????? ????? ??10? ????? ????4 1?0-0 0???? 10111 11211 41000 01010 00001 ????? ????? ????? ????? ?00?1 11??? 21?0? 0???2 00001 00?00 00011 21112 11010 10201 01101 30100 ????- 0?0?0 1???? ????? ????? ????? ??1?1 12001 00000 10000 01020 0???? ????? ????? ????? ?? 1010? ????? ?1??0 22000 0???? 31000 11000 00001 Jeholornis 00000 00010 0???? ??202 2??0? ??0-1 ?1103 00101 01200 100?0 012?0 0???? ????? ????? ????? ?? 12??0 ????? 11?10 0???? ?1?1? ????? 1???? ????? ????1 00001 ????? ????? ????? ????? ????? ????1 Nothronychus 10— 2000- 2–11 12110 0001- 00010 ??0?? ??1?? ????? ????? ????? ????? ????? ????? ????? ????? ????? 1???? ?2?23 11112 0?000 10011 11210 41111 22101 ????? ????? ????? 01111 0001? ???01 1010? ????? 00012 00011 0?101 00120 0-012 21000 210-1 0???? ????0 0???0 ????? ????? ????? ???0? ??110 1(1,5)10? ????? ?0000 10020 11021 2???? ????? ????? ????? ?? ?1?2? ????? ????? ????? ????? ?110? ????? ????0 10001 00100 0???? ????? ????? ?031? 0???? ????0 Juravenator 0000? ????? ????? ????? ????? ????? ????? ?? 0—0 00000 00011 00002 -0001 01000 0???? ????? Ornitholestes ?0000 10000 0???? ????? ????? ????? ????? ????0 00201 10000 0000? 0???0 0001- 000?? ????? ??1?? ????? 1000? 01011 0000? ?1001 00000 000?0 00??1 1???? ????? ?010 0 00000 0—- -11– 21000 00??? 00000 ?0??0 0???0 ???1? 0??0? 001?? ????? ????0 ????1 0000? ??000 00001 ????? ????? ????? ????? 00010 00000 ?000? 00000 00100 ??0?? ???0? ??110 ????0 0000? 100?0 01020 0???? ????? ????? ????? ?? 12101 00012 ??1?1 0???? ????? ????? ????? ????? ????? 00001 00??? ?0022 21002 01000 ?0100 0???? Longisquama ?010? ????? ??0?0 010?0 0???? ????? ????? ????? ?? ?—0 ????? 0??11 ??01? ?10?? 011?? 0???? ????? Ornithosuchus ????0 ????0 1???? ????? ????? ????? ????? ????? 00??2 2000- 2?11? 1???0 0?01- ????? ????? ????? ????1 10100 01010 00100 00110 00011 00000 00?00 0???? ????? ????? ??000 0—- -0210 11010 00000 ?0100 10000 001?? ?100? 00??? ?00?? ????? ????0

OM66_Text.indd 73 3/31/09 5:55:22 PM 74 ORNITHOLOGICAL MONOGRAPHS NO. 66

00000 00000 ?000? 00000 00100 10000 00010 ??1?0 ??101 00120 0-??? 21000 2???? ???0? ????1 00000 0110? 01?01 00100 0???0 —– -00– 21000 00000 10000 11030 2???? ????? ????? ????? ?? 00001 0000? 1?001 00020 11001 11000 00100 00000 Rahonavis 30000 1000? 11100 01010 0?000 00100 00000 000?? 00 ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? Oviraptor ????? ????? ????? ????? ???1? ??1?0 11101 02?23 ????? 1?0?? ?1?11 ?100? ??0?? 0?00? 1???? 1???? 12112 0???? ????? ????? 310?? ????? ????0 ????? ????? 1?0?1 ????1 ???0? ????? ????? ????? ????1 11?– ??102 00110 0-002 01?0? ?1301 01102 30100 00100 —– —– 11211 00001 0?011 1???? ????? ????? ????? 10000 01031 2???? ????? ????? ????? ?? ????? ????? ????? ????1 ????0 0???? ????? ????? Sapeornis ???0? ?0021 ????? ????? ????? ????? ????? ????? ????? 0???? ????? ????? ????? ????? ?? ????? 10100 01010 00012 -1001 0?1?? 011?1 ????? ?0?00 10000 1???? ????? ????? ????? ????? ????0 Patagopteryx 01202 —– 211-1 10000 0001- 0001? ??0?? ??1?0 ????? ????? ????? ??01? ????1 1???? ????1 ?11?1 12-11 14?0? ?1??4 120-0 1?000 0—- -1211 11011 21010 ????1 1???? ????? 11??? 0???1 11?11 21??? ?0??? ????? 00000 00002 00100 0?120 0-0?2 21001 210-1 01003 ????? ????? ????? 00??1 1??11 10111 14001 00004 00101 00200 10010 11020 2???? ????? ????? ????? 1???0 ???0? 2???1 1???? 41101 22101 00??0 00012 ?? 0011? ?01?? 0-211 21100 00100 01112 -0101 00200 Saurornithoides 10030 21030 1???? ????? ????? ????? ?? ????? 00101 00111 10012 -20?1 01000 0???? ????? Pelecanimimus ?0?00 10001 0???? ?111? ???1? 100?? ????? ?01?0 ????? 10001 00011 00010 00101 ????? 0???? ????? 00101 111?1 ?111? 0010? 01??0 ??1?? ????? ????? ????? ??0?? ????? ????? ????? 1???? ????? ????0 1???? 0???? ??1?0 0???? ????? ????? ????? ????? 00212 2220- ?111? 1??0? 0???? ?00?1 ??1?? ????? ????? ????? ????? ????? ????? ???0? ??201 11101 5???? ?0??? ????? ????? 10000 0???? ?100? ?1000 311?? ????? ????0 01121 0???? ????? ????? ????? 00??? ????? ???00 0???? ????? ????? ????? ????? ?? ????? ????? ????? ????? ????? ????? ????? ????? ?? Scleromochlus Postosuchus ????1 00000 00010 0020? ????? 000?? 0???? ????2 ????1 00000 00011 00100 00100 00011 00000 01002 ?2-00 10000 0011? ????? ????? ????? ????? ??1?0 02-00 00000 00110 00110 00101 000?1 00100 ?0000 00??? ???0? ????? ?1000 00000 000?0 ??20? ???0? 00000 00000 00001 00000 00100 10000 00100 10100 ?1?0? ????1 01??0 00000 0—- -11– 01010 00000 01100 00121 0?100 00000 0—- -11– 20000 00000 00?00 00020 00000 00020 0-000 11000 000-? 00103 00001 00000 00110 00000 0-001 11000 20100 00103 00100 0000? ??000 01010 0???? ??0?? 00000 00??? 00100 10001 11100 00020 0?000 002?? 00000 00000 ?0 00 Shuvuuia Proterosuchus ????? 00101 0001? ?0012 -0001 012– 13-01 10110 ????? 00000 00000 00100 00000 00011 00000 00000 01000 10001 0??20 ?3— 1??0? 000?? ???01 ????0 00000 10000 00000 00000 00000 00000 00000 000-0 00??? ??01- ??00? 11000 00000 00010 ??010 ??111 00000 00000 ?000? 00000 0001- 00010 00300 00000 12001 10?14 11110 01100 22000 031– 31000 11000 01000 -0000 000-0 00000 0—- -00– 00000 00000 00001 00000 00010 000?? ??212 20100 000-1 01102 00000 00020 00000 00020 0-010 00000 000-0 00?03 30101 01100 10000 01220 0???? ????? ????? ????? 00100 00000 00000 00000 0?000 00000 00000 00??? ?? ?0 Sinornis Protopteryx ????? 11100 110?? ??012 -???1 ????? 1???? ????? 12000 1110? 11??? ??012 -???1 011?? ????0 0??01 ????1 10001 1?1?? ????? ????? ????? ????? ????0 0???? ????? 1???? ????? ????? ????? ????? ????0 1020- 1020- 2000- 21-11 1???? ??01- 0?01? ???1? ??111 200?- 21-11 1???? ??01- ????? ????? ??1?? 1??0? ????4 13?0? 01??4 12110 12101 20011 13222 41010 120-0 1?100 21011 13222 41011 22101 11?10 00112 22101 01110 00112 00102 ?0120 22212 21000

OM66_Text.indd 74 3/31/09 5:55:22 PM origin of birds 75

210-1 01102 -2101 00200 10020 11030 2???? ????? Sphenosuchus ????? ????? ?? ????1 00000 00010 00001 00001 00000 00000 01112 Sinornithoides 02-00 10001 00110 00011 01101 00011 01101 11100 00100 00000 00000 10000 00000 00000 01010 00100 ????? 00101 00111 ?0002 -101? ????? ?00?? ??00? 00000 0000? ????0 ?0100 0—- -01– 10000 02100 ?1-00 1000? ??12? ????? ????? ????? ????? ??1?0 00001 00001 00??? ????? ????? ????? ????? ????? 00001 11110 01110 0000? 01?00 ????0 ???10 ??1?1 ????0 ?000? ????0 010?0 0?010 00??? ????? ????? ?? 11??? 00?2? ?1110 01000 0—- -10– 11000 01000 00001 0001? 0?102 0?020 ??112 11000 0030? 11101 Syntarsus 311?0 00100 10000 01121 0???? ????? ????? ????? ?? ????? 00100 00011 00000 00001 00000 00000 00000 Sinornithomimus 00000 00000 0?1?? ?111? 01101 00001 00?00 10100 00200 20000 00000 00000 00100 00000 00210 01100 ????? 00001 00011 10010 01001 01000 00000 010?1 01110 00021 10100 00000 0—- -1200 10000 00000 00000 10000 0???? ????? 01??1 100?? ???01 1?1?1 00001 00000 00011 00000 20-02 11000 000-0 01100 11— —– —– 10000 00000 00010 ??201 10110 1?10? 10000 00200 10010 11020 0???? ????? ????? ????? ?? 00?23 00101 0?000 0—- -11– 11000 01000 00001 00021 00000 00101 ??002 11000 100-? 01100 20100 Terrestrisuchus 00100 10000 01220 3???? ????? ????? ????? ?? ????1 ????0 00010 0000? ?0001 00011 00000 01112 Sinornithosaurus 02-00 00000 0?110 0100? 10?1? ?0011 11??1 ????? ????? 00000 01011 10000 11001 01000 11100 10001 00??0 00000 ??00? 00000 00000 00000 0131? ??1?0 11-00 11111 1??1? ????? ????? 0???? ????? ????0 0???? 0???0 00100 01000 00000 001– 10000 01100 00201 (1,2)0001 00001 00000 00100 00100 ??11? 00001 00001 00111 00020 0-012 11000 000-0 00003 ????? 11??? 1???2 1?112 0?011 10101 11210 411?1 00100 10001 11100 01010 0?000 00111 00000 11110 00000 00012 ??101 0???? ???11 21001 21301 01000 00 11103 011?0 00100 10000 01121 0???? ????? ????? Troodon ????? ?? ????? ??1?? 0??11 ?0012 -1001 01000 ????? ?11?? Sinosauropteryx ?0000 00001 0???0 ????? 11101 10011 11111 1011? ????? 00000 00011 00000 10001 00000 0??00 01??1 00001 11111 01000 0010? 01??? ????? ???11 ??111 00000 1000? 0???? ????? ????? ?00?? ????? ????0 11101 00113 0?1?0 010?? ????? ????? ????? ????? 00201 10000 00000 0???? ??01- 0???? ???00 ??100 ????? 0000? ??111 00??? ????? 01?00 102?1 11101 01?0? 00?1? ?0100 00000 0—- -11– 11000 00000 3110? ?0200 10000 01121 0???? ????? ????? ????? ?? 00001 00000 ?0011 01000 12012 11000 1010? 01100 Tyrannosaurus 20000 00000 10000 01020 0???? ????? ????? ????? ?? ????? 00000 00011 10102 -0111 00000 00000 01001 00100 01101 01120 00111 01111 000?1 00?01 10000 Sinovenator 00010 00000 00000 00000 00100 11000 01301 00100 ????? 00001 01?11 10012 -???1 ????? 0???? ?1??? 01100 02022 01100 00000 —– -1200 21000 00000 ????? 10001 1??2? ????? 11111 000?1 11?01 21100 00001 00000 00010 01002 10002 11000 10101 01100 00201 111?0 ?111? 0000? 11??? ????0 ???1? ??11? 30000 00000 10000 01120 0???? ????? ????? ????? ?? 01101 00?12 01110 0???? ????? ????? 310?1 11010 Unenlagia 00?00 ????? ????? ?0021 21002 21000 21301 11101 31100 01000 10000 01021 0???? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ?? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ????? ??1?? ?11?1 121?3 Sinraptor ????? ????? ????? ????? 3101? ????? ????0 00001 ????? 00000 00011 11100 00111 00000 00000 00001 00??? ?0121 21001 01000 21301 11101 301?? ????? 00000 00000 01100 01110 01101 00001 00?00 ?00-0 ????? ????? ????? ????? ????? ????? ?? 00000 00000 00001 10000 00100 110?0 00111 00100 Velociraptor 02100 02022 0?100 ?0000 21000 01??? 3100? ????? ????? ????? ???0? ?1002 10002 11000 10100 01100 ????? 00000 00011 10002 -1001 01000 01000 10001 20010 00000 10000 01020 0???? ????? ????? ????? ?? 00000 11110 01120 01111 11101 001?1 00?01 11100

OM66_Text.indd 75 3/31/09 5:55:23 PM 76 ORNITHOLOGICAL MONOGRAPHS NO. 66

00000 00000 00000 10000 00100 10100 00311 10110 Conchoraptor 11101 1?123 11112 01011 10101 11200 31001 11010 Character 103: See Elzanowski (1999). 00000 00001 00101 00011 21102 21000 20301 11101 31100 00000 10000 01021 0???? ????? ????? ????? ?? Character 151: Clark et al. (2002) did not code the ac­ romion as laterally everted, whereas Marya´nska et Yanornis al. (2002) did. Osmólska et al. (2004) stated that the ????? 11100 11010 00012 -1001 1—- 1???2 ????1 acromion of oviraptorosaurs is distinctly laterally ?2-00 10000 1???? ????? ????? ????? ????? ????0 everted in contrast to that of most theropods. This 00202 200?- 21111 1?000 0001- 0???? ????? ??1?? state is clearly illustrated by the scapulocoracoid of ????? ?1??4 1???? 2???? 22011 12211 41011 ?2101 Ingenia (fig. 8.2G, H of Osmólska et al. 2004). 10?10 01112 ???11 ??1?? ???1? 21000 00??? 0???? Deinonychus ????1 00200 10020 21030 2???? ????? ????? ????? ?? Character 20: Currie (1995:580) provides the justi­ Appendix 5: Notes Accompanying Matrix fication for this coding. on Scoring Decisions for Certain Taxa Dibrothosuchus Alligator Character 163: See fig. 10C of Wu and Chatterjee (1993). Character 125: Contrast figure 3 of Mook (1921) with plate 18 of Madsen (1993). Erlikosaurus Character 197: The condition in Alligator is simi­ Character 103: See Elzanowski (1999). lar to that in Marasuchus (see fig. 9A of Sereno and Erpetosuchus Arcucci 1994). Character 148: Although the clavicles are not pre­ Alxasaurus served with the skeleton, an attachment for them Character 129: Russell and Dong (1993b) present is present on the scapula according to Benton and no evidence to show that all the presacral verte­ Walker (2002:33). brae were pneumatized. Erythrosuchus Archaeopteryx Character 129: Gower (2001) detailed the pres­ Character 96: See Martin and Stewart (1999), ence of pneumatic dorsals in Erythrosuchus but whose account we confirmed by personal obser­ demurred in his 2003 monograph on this taxon. vation of casts at UKMNH, contrary to the de­ Nonetheless, his earlier arguments were compel­ scription of Elzanowski and Wellnhofer (1996). ling and are followed here. Character 221: Following Mayr et al. (2005, Euparkeria 2007). Character 127: Ewer (1965:429) discusses “central Avimimus excavations” in the vertebral column, which she suggests are related to the pneumatic excavations Character 40: See figure 1a of Kurzanov (1985). of the vertebrae in birds and some other archo­ Character 80: Kurzanov (1985) described teeth in saurs. Because these are not clearly illustrated the type specimen, but these are really denticles and their nature remains uncertain (see also formed by the crenulate margin of the premaxil­ Gower 2001), we have coded this character and lae (Vickers-Rich et al. 2002). character 129 as unknown. Citipati Gallimimus Character 103: See Elzanowski (1999). Character 103: See Elzanowski (1999) and Hurum (2001). Compsognathus Gansus Character 20: Although it is not preserved as a sep­ arate element, a sutural facet on the anteroventral Character 140: Costal facets do not appear to oc­ edge of the left frontal probably indicates that the cur on the sternum (see fig. S4 of You et al. 2006, prefrontal was present in life (Ostrom 1978:83). supplementary material).

OM66_Text.indd 76 3/31/09 5:55:23 PM origin of birds 77

Harpymimus as homologous with the acrocoracoid process, but the structure is not appreciably different from that Character 103: See Elzanowski (1999). found in some theropods and in Archaeopteryx, in Herrerasaurus which taxa an acrocoracoid is absent. Character 186: Contrary to Paul (2002), the pubis, Scleromochlus despite being somewhat posteriorly reflected, is not actually retroverted as it is in maniraptorans Character 154: See figure 2a of Benton (1999), and birds. which shows a cast of the dorsal slab of BMNH R3146, clearly indicating that the scapula is po­ Heyuannia sitioned parallel to the vertebral column, as also Character 103: See Elzanowski (1999). reported by Martin (1983, 2004). Benton’s figure 14, however, shows Scleromochlus restored with Hongshanornis the scapula far down the ribcage and not parallel Character 140: Costal facets do not appear to oc­ to the vertebrae. cur on the sternum (see fig. 4 of Zhou and Zhang Character 165: The orientation of the scapula neces­ 2005). sitates a lateral orientation for the glenoid fossa. Longisquama Shuvuuia Character 14: Our examination of high-quality Character 20: We concur with Chiappe et al. stereo photographs and casts of the Longisquama (2002), rather than with Sereno (2001), on the in­ material available at UKMNH confirmed the terpretation of the preorbital region of Shuvuuia. presence of the antorbital fenestra (see Fig. 15). Sinornis Character 15: Martin (2004, fig. 4D) restored the skull of Longsiquama with two accessory antor­ Character 14: Martin and Zhou (1997) provided a bital fenestrae, of which we could confirm the reconstruction of the antorbital cavity for the re­ presence of one (see Fig. 15). ferred specimen (IVPP V 9769) that we could not confirm from observation of the casts available Ornithosuchus to us. A clear antorbital cavity cannot be distin­ Character 103: See Walker (1964:76, 115); see also guished. Sereno et al. (2002:189) also note difficul­ Walker (1961) and Romer (1956). ties in deciphering the structure of the antorbital cavity. Patagopteryx Character 15: Martin and Zhou (1997) restored the Character 141: The sternum is not completely antorbital fossa as being pierced by an accessory known, but no evidence indicates a suture or that maxillary fenestra, but we could not confirm this the sternal plates were unfused. character from personal observation of the speci­ Postosuchus men or from other descriptive accounts in the literature. Sereno et al. (2002:189) suggested the Character 15: See figure 3b of Chatterjee (1985). possible presence of this accessory fenestra, but Character 129: See Gower (2001) and Chatterjee the antorbital region is too poorly preserved for (1985:417–418, fig. 12). certainty. Sanz et al. (1997) noted the absence of such a fenestra in the skull of a hatchling enantio­ Protopteryx rnithine from Spain. Character 208: Although the tibia is referred to as Character 22: Martin and Zhou (1997) asserted a tibiotarsus by Zhang and Zhou (2000), the prox­ that a “T-shaped” lacrimal is present in the re­ imal tarsals are not fused to the tibia, as is appar­ ferred specimen. We could not confirm this char­ ent from the description of the tarsus (Zhang and acter upon inspection of the casts available to us, Zhou 2000:1956–1957), and the term “tibiotarsus” and Sereno et al. (2002) made no mention of a lac­ is therefore inappropriate. rimal in their description of the cranial anatomy Sapeornis of Sinornis. Character 160: Zhou and Zhang (2003a:736) regard Character 34: Martin and Zhou (1997) stated the “biceps tubercle” of the coracoid in Sapeornis that “there is a well preserved quadrate . . . lying

OM66_Text.indd 77 3/31/09 5:55:23 PM 78 ORNITHOLOGICAL MONOGRAPHS NO. 66

disarticulated and behind one skull. . . ,” but in­ and Sphenosuchus. In the former, we consider it spection of the skull of the referred specimen fails a foramen transmitting the “temporo-orbital” to corroborate this assertion, a conclusion also (= stapedial) artery, but in Sphenosuchus it is prob­ reached by Sereno et al. (2002). ably a pneumatic feature. Character 127: Although the pleurocoels are Character 71: In contrast to the condition in Sphe- poorly preserved in the holotypic specimen, an nosuchus, clear evidence shows that the metotic isolated dorsal with a distinct pleuorcoel is pres­ fissure had been subdivided by a prevagal com­ ent in the specimen examined. missure. Sinornithoides Character 141: Walker (1990:64, and unpublished correspondence dated 29 October 1995) disputed Character 20: See Currie and Dong (2001:1755). the identification of the ossified median element Character 208: Although Sinornithoides is often identified by Crush (1984, fig. 7B) as a sternum. described as possessing a fully avian tibiotarsus, He argued that it was in fact an interclavicle. only the proximal tarsals are in fact fused. His reasoning, though perhaps applicable to a similar element in Sphenosuchus, certainly does Sinornithosaurus not hold in the case of Terrestrisuchus, in which Character 33: The quadratojugal has drifted away Crush (1984) clearly identified another element as from the quadrate in the skull of IVPP V 12811 the interclavicle and a second median element in (fig. 2 of Xu and Wu 2001), which suggests that it the pectoral girdle as the sternum. Given that the was only ligamentously attached. coracoids articulate directly with this second me­ dian element, we conclude that Crush (1984) was Character 42: See Xu and Wu (2001:1750) for the correct and that this element is a sternum. coding of this character. Character 143: Unpublished correspondence Character 185: Some medial closure of the acetab­ from A. D. Walker dated 29 October 1995 notes ulum was noted by Xu et al. (1999b) and is appar­ that examination of the element Crush (1984) ent in their fig. 4e. identified as a sternum failed to reveal any trace Sphenosuchus of rib facets. Character 112: See Walker (1990:53 and fig. 34g). Troodon Character 129: No pneumatic features occur in Character 20: See Gauthier (1986) for commen­ any of the vertebrae figured by Walker (1990). tary on the structures Currie (1985) identified as prefrontals. Character 166: See Sereno (1991:27). Yanornis Terrestrisuchus Character 149: Zhou and Zhang (2001) describe Character 37: Crush (1984) figured what clearly the furcula as “U-shaped,” but their figure 2 appears to be a pneumatic foramen perforating (p. 1259) clearly illustrates a furcula that does not the quadrate, though no clear reference appears appreciably differ from that of Archaeopteryx or in the text to the presence or absence of this char­ Confuciusornis, except for the presence of a small acter. A similar foramen is found in Dibrothosuchus hypocleidium.

OM66_Text.indd 78 3/31/09 5:55:24 PM